Full text of "The Wilson bulletin"

See other formats

HARVARD UNIVERSITY

Library of the
Museum of
Comparative Zoology

Hie WIson Bulletin

PUBLISHED BY THE WILSON ORNITHOLOGICAL SOCIETY

VOLUME 90 1978 QUARTERLY

REVIEW editor: ROBERT RAIKOW

COLOR PLATE EDITOR: WILLIAM A. LUNK

EDITORIAL ASSISTANTS: BETTE J. SCIIARDIEN

C. DWIGHT COOLEY
PATRICIA ILAMEY
MARTHA B. HAYS
GARY L. MILLER
RENNE R. LOHOEFENER

I’n sidrnt — Douglas A. .Iam«-s, I )rparliiu‘nt (»f Zoology, I niversity of Arkansas, Fayetteville,
Arkansas 72703.

First Vi«»*-l*resi(l«‘nl (ieorge A. Hall. Department of Chemistry, West Virginia Univer-
sity. M(»rganl«»wn, \\ Csl \ irginia 20.300.

.'^eeornl Vice-President — Ahhot .S. (iaunt, Department of Zoology, Ohio I^tate University,
(adumhus, Ohio 43210,

Kditor — .l«‘roim* A. Jackson, Department of Hiol(»gieal .Sciences, P.O. Drawer Z. Mississippi
.Statj* University, Mississippi .State, Mississippi .39702. (.See Ornithological News, p.
1.3«).

.S«-eretarv — Curtis .S. Adkisson, Department of Biology. Virginia Polytechnic Institute and
.State University, Blackshurg. \ irginia 24001.

Treasurer — Krnest K. Hoover, 1044 Webster St., N.W., Grand Rapids, Michigan 49504.

Fleeted (aumcil Memhers — James K. Karr (term exjjires 1979); Clait E. Braun (term
expires 1980); Sidney A. (iauthreaux. Jr. (term expires 1981).

I)\TKS OF ISSUE OF VOLUME 90

OF I'llK W ILSON BULLETIN

\o. 1—19 April 1978
NO. 2 — 16 August 1978
NO. :3 — 21 Novemher 1978
NO. 1 1.5 FERRUARY 1979

CONTENTS OF VOLUME 90

NUMBER 1

owTH AND SURVIVAL OF YOUNG FLORIDA SCRUB JAYS Glen E. If oolfenden

RTICAL DISTRIBUTION OF BIRDS IN A LOUISIANA BOTTOMLAND HARDWOOD FOREST

James G. Dickson and Robert E. Noble

rRICULTURAL IMPACT OF A WINTER POPULATION OF BLACKBIRDS AND STARLINGS

Richard A. Dolbeer, Paul P. Woronecki, Allen R. Stickley, Jr., and Stephen B. White

;eeding behavior of the Louisiana heron James A. Rodgers, Jr.

ATUS AND numerical FLUCTUATIONS OF SOME NORTH AMERICAN WADERS ALONG THE

SURINAM COAST ^rie L. Spaans

;eding of nestling and fledgling eastern bluebirds Benedict C. Pinkowski

FFERENTIAL USE OF FRESH WATER ENVIRONMENTS BY WINTERING WATERFOWL OF COASTAL

xexas Donald H. White and Douglas James

■PRODUCTIVE SUCCESS AND FORAGING BEHAVIOR OF THE OSPREY AT SEAHORSE KEY, I'LORIDA

Robert C. Szaro

SNERAL NOTES

CHANGING AVIAN COMMUNITY STRUCTURE DURING EARLY POST-FIRE SUCCESSION IN THE

SIERRA NEVADA Garl E. Bock, Martin Raphael, and Jane H. Bock

NOTES ON THE DISTRIBUTION OF BIRDS IN SONORA, MEXICO

Stephen M. Russell and Donald W . Lamm

EGG CARRYING BY WOOD DUCK

Robert W. Strader, Richard Di Giulio, and Robert B. Hamilton

EVIDENCE OF BROOD ADOPTION BY RUFFED GROUSE Stephen J. MaXSOn

MARSH HAWKS FOLLOW' HUNTING RED FOX LeRoy ff . Bandy and Barbara Bandy

PREDATION ECOLOGY OF COEXISTING GREAT HORNED AND BARN OWLS Seri G. Rudolph

HOST RECORDS FOR THE STRIPED CUCKOO FROM COSTA RICA

Lloyd F. Kiff and Andrew ff illiams

ANT-FOLLOWING BIRDS IN SOUTH AMERICAN SUBTROPICAL FORESTS

Michael Gochfeld and Guy Tudor

FISHING BEHAVIOR OF BLACK AND TURKEY VULTURES

Jerome A. Jackson, Irvine D. Prather, Richard N. Conner, and Sheila Parness Gaby
A NEW HYBRID WARBLER COMBINATION Richard C. Banks and James Baird

RNITIIOLOGICAL LITERATURE

RNITHOLOGICAL NEWS

EQUESTS FOR ASSISTANCE

NUMBER 2

REVISION OF THE MEXICAN PICULUS (PICIDAE) COMPLEX Luis F. Baptista

ISTRIBUTION, DENSITY, AND PRODUCTIVITY OF ACCIPITER HAWKS BREEDING IN OREGON

Richard T. Reynolds and Howard M. Wight

OCIAL AND FORAGING BEHAVIOR OF WARBLERS WINTERING IN PUERTO RICAN COASTAL SCRUB

William I*ost

DE RESIDUES AND EGGSHELL THINNING IN LOGGERHEAD SHRIKES

William L. Anderson and Ronald E. Duzan

112

119

123

131

132

133

134

138

139

141

143

145

157

118

159

182

197

215

n<^^. s|-K( IKS I SKI) in IlIKDS IN I «)(,(. KD \M) I NLoGOKI) MIXKD-CO.MI KHOL S FORKSTS

Kathleen E. Franzreb 221 ;

1)01 III riiHMoDKDNKSS IN I’l RiM.K MARTINS IN TKXAS Charles R. Brown 239 t

H)oi) OF NvsTi.iNi; iTRi’i.K MARiiNs Helene Walsh 248.

RKl’Kom ( TloN AND NKST SITK SKI.F.CTION RY RKD-WINGKI) RLACKRIRDS IN NORTH LOUISIANA

Bryan T. Brown and John W. Goertz 261
TIIK RUFors-coLLARKi) SPARROW AS A HOST OF TiiK SHINY cowRiRD Rosendo M. Fmga 271

(.KNKHAI. NOTKS

SKXL’AL SIMILARITY OF RKD-IIKADKI) WOODPFCKKRS AND POSSIRLE EXPLANATIONS BASED

ON FALL TERRITORIAL REH AMOR — - Lawrence Kilham 285

NOTES ON THE COURTSHIP REHAVIOR OF RROW N-C APPEI) ROSY FINCHES

Paul Hendricks 285

EFFECTS OF NEST REMOA AL ON STARLING POPULATIONS

// W' Heusmann and Robert Bellville 287

UNUSUAL INCUBATION REHAMOR IN RORAVHiTE George A. Hurst 290

A CATTLE EGRET-DEER MUTUALISM Marc R. Halley and Wayne D. Lord 291

A TEST OF SIGNIFICANCE FOR MAYFIELD’s METHOD OF CALCULATING NEST SUCCESS

Douglas D. Dow 291

MIRROR IM AGE AERSI S CONSPECIFIC STIMULATION IN ADULT MATE ZEBRA FINCHES

Michael J. Ryan 295

PROTOCAU.irilORA INFESTATION IN GREAT HORNED OWLS Robert T. Bohm 297

TERRITORIAL DEFENSE OF A NECTAR SOURCE BY A PALM AVARBLER

Joseph M. Wunderle, Jr. 297

RING-RILLED (H LL PAIR AATTH 2 NESTS William E. Southern 299

CLUTCH SIZE AND NEST PLACEMENT OF THE PIED-BILLED GREBE IN MANITOBA

Spencer G. Sealy 301

NEST PLACEMENT IN SAGE THRASHERS Terrell D. G. Rich 303

GREAT BLACK-RACKED GULLS BREEDING IN SALT MARSH IN NEAV JERSEY

Joanna Burger

president’s PA(.E . ...

ORNITIHH.«)GICAL NEAAS

H).NSKRA ATION ( OMMITTEE REPORT

«)RNITHOLOGICAL UTERMURE

304

306

308

309
322

NUMBER 3

Al TI M.N BIRD CASI AI.IH.S AT A NORTHAVEST FLORIDA TA TOAVER: 1973-1975

Robert L. Crawford 335^

WHITE PELICAN PRODUCTION AND SURVIVAL OF YOUNG AT CHASE LAKE NATIONAL WILDLIFE

REFUGE, NORTH DAKOTA Robert F. Johnsoji, Jr. and Norman F. Sloan 346

EGG VOLUME AS A PREDICTOR OF HATCHLING WEIGHT IN THE BROWN-HEADED COWBIRD

I al Nolan Jr. and Charles F. Thompson 353

BEHAVIOR AND SEX ROLES OF NESTING ANHINGAS AT SAN BLAS, MEXICO

Joanna Burger, Lynne M. Miller, and D. Caldwell Hahn 359

POST-FLEDGING BEHAVIOR OF PURPLE MARTINS Charles R. Brown 376

NESTING ECOLOGY OF THE PLAIN CHACHALACA IN SOUTH TEXAS

Wayne R. Marion and Raymond J. Fleetwood 386

SPATIAL RELATIONSHIPS IN PERCHING BARN AND CLIFF SWALLOWS Anne E. HuttOn 396

POPULATIONS OF BAY-BREASTED AND CAPE MAY WARBLERS DURING AN OUTBREAK

OF THE SPRUCE BUDWORM Douglass H. Morse 404

AGE AND FORAGING ABILITY RELATIONSHIPS OF OLIVACEOUS CORMORANTS

Michael L. Morrison, R. Douglas Slack, and Edwin Shanley, Jr. 414

ANALYSIS OF ROOSTING COUNTS AS AN INDEX TO WOOD DUCK POPULATION SIZE

Delbert E. Parr and M. Douglas Scott 423

GENERAL NOTES

BLACK SKIMMER ABUNDANCE ON THE LOUISIANA-MISSISSIPPI-ALABAM A COAST

John W . Portnoy 438

KiLLDEER BREEDING DENSITIES Terrence R. Mace 442

BROWN PELICAN RESTOCKING EFFORTS IN LOUISIANA

Stephen A. Nesbitt, Lovett E. Williams, Jr., lAtrry McNease, and Ted Joanen 443

NOTES ON 2 SPECIES OF BIRDS PREVIOUSLY UNREPORTED FROM PERU

Dan A. Tallman, Theodore A. Parker, III, Gary D. Lester, and R. A. Hughes 445

RESPONSES OF BIRDS TO A SNOWSTORM IN THE ANDES OF SOUTHERN PERU

John P. O’Neill and Theodore A. Parker, HI 446

CANNIBALISM BY AN ADULT GREAT HORNED OWL

J. B. Millard, T. H. Craig, and O. D. Markham 449

CACHING BEHAVIOR OF SCREECH OWLS IN INDIANA James B. Cope and John C. Barber 450

ATTACKS ON RED-HEADED WOODPECKERS BY FLYCATCHERS Roland R. Roth 450

AN ANALYSIS OF GILA WOODPECKER VOCALIZATIONS Gene L. Brenowitz 451

AN AGGRESSIVE ENCOUNTER BETWEEN A PINTAIL WITH A BROOD AND A FRANKLIN GULL

George Hochbaum and Garth Ball 455

CANADA GOOSE-GREAT BLUE HERON-GREAT HORNED OWL NESTING ASSOCIATIONS

Richard L. Knight and Albert H . Erickson 455

GIANT CANADA GOOSE INCUBATES EGGLESS NEST Conrad A. Fjetland 456

NESTING SUCCESS AND NEST SITE SELECTION OF RED-WINGED BLACKBIRDS

IN A FRESHWATER SWAMP Brent Ortego and Robert B. Hamilton 457

EXTREME NESTING DATES FOR THE MOURNING DOVE IN CENTRAL ILLINOIS L. Barrie Hunt 458

A VOLUMETRIC ANALYSIS OF SHARP-TAILED GROUSE SPERM IN RELATION TO DANCING

GROUND SIZE AND ORGANIZATION Wayne M. Nitchuk and Roger M. Evans 460

PRESIDENTS PAGE 463

FIFTY YEARS OF THE WILSON ORNITHOLOGICAL SOCIETY Maurice BrOoks 464

oH M I H« »l ( il.K M IIIKHMIHK

CiiNslin I|n\ \M» HV-I \\\> <»K TIIK WIl.SON ( )H MTHOI.OCilCAI. SOCIKTY

OKMrimi.ni.M M. >K\\S -

NIIMFIKR 4

MS M\II KMION UK iiix'fkk's - (ieoffie M. Sultan and David F. Parmelee

IIIKII I.IKK. \T CM’K (KU/IKK, KUSS ISLAND

David G. Ainley, Robert C. Wood, and William J. L. Sladen

.s|/K, KU<ID TM'K. XND KuK ACTING SI TKS UF RKD-WI NGKD BLACKBIRDS — Stephen W. Wllson

WINTKK TKRKITURIAL WD KuK ACiINfi BKII AVIOB OF RP:D-HEADED WOODPECKERS IN FLORIDA

Debra Moskovits

TVKSAL col. OR OF AMERICAN COOTS IN RELATION TO AGE Richard D. Cruwford

NFSTINI, BKIIAAloR AND AFFINITIES OF MONK PARAKEETS OF SOUTHERN BUENOS AIRES

PROAINCE, ar(;entina Philip S. Humphrey and Roger Tory Peterson

M<*RPII0L0(.Y OF THE LARYNX OF CORVVS B RACHY RHY \CHOS (PASSERIFORMES: CORVIDAe)

Walter J. Bock

IIAIUTAT I SE BA A ELLOW-RI M PED WARBLERS AT THE NORTHERN EXTREMITIES OF THEIR

WINTER RAN(,E Kenneth J. Wllz and Vincent Giampa

HABITAT SHIFT AND ROADSIDE MORTALITY OF SCARLP:T TANAGERS DURING A COLD WET NEW
EN(.LANi) SPRING David C. Zumeta and Richard T. Holmes

loMMI NITA KCoLoGA OF THE HELMINTH PARASITE:S OF THE BROWN PELICAN

Stephen R. Humphrey. Gharles H. Courtney, and Donald J. Forrester

NEST-SiTE SELECTION OF AVILLETS IN A NEW JERSEY SALT MARSH

Joanna Burger and Joseph Shisler

UR(. ANUCIILORINE RESIDUES AND EGGSHELL THINNING IN WOOD STORKS AND ANHINGAS

Harry M. Ohlendorj. Erwin E. Klaas and T. Earl Kaiser
II Aim AT SELECIION BA BREEDING RED-AVINGED BLACKBIRDS Peter H. AlberS

GENERAL NuIE>

ANUIIIER COLONA OF HIE (,l ADELOl PE HOUSE AAREN J on C. BarloW

PEMH.IDE LEAELS AND SHELL THICKNESS OF COMMON LOON EGGS IN NEAV HAMPSHIRE

Scott A. Sutcliffe

DE»LINE.s IN ENA IRON MEN ! AL POLLUT ANTS IN OLIVACEOUS CORMORANT EGGS FROM TEXAS.

1970 1977 Michael L. Morrison, R. Douglas Slack, and Edwin Shanley, Jr.

1IRKEA AILIIRE E<.(. SHELL TIIINNINT, IN CALIFORNIA. FLORIDA. AND TEXAS

Sanford R. W ilbur

AN EXPERIMENIAI. ANAI.AsIS OF HIE INTERRELATIONSHIP BETAVEEN NEST DENSITY AND

pREDAiioN IN OLD-MELD HABITATS Bradley 4/. Gottfried

CANADA GOOSE TAKES OVER MALLARD NEST . Thomas N. Mather

NoIM on MiOD HABITS OF THE PLAIN CHACHALACA FROM THE LOAVER RIO GRANDE VALLEY
Zdn D. (.hristensen. Danny B. Pence, and Gretchen Scott

468

474

478

479

492

511

521

536

544

553

566

575

587

599

608

619

635

637

641

642

643

646

647

HERRING GULLS STEALING PREY FROM PARASITIC J AEGERS R. 1. G. Morrison

THE USE OF FEEDING AREAS OUTSIDE OF THE TERRITORITY OF BREEDING BLACK

OYSTERCATCHERS E. B. Hartwick

SCREECH OWL PREDATION ON A COMMON FLICKER NEST Mary C. Landin

RED BOBWHiTES IN OKLAHOMA Jack D. Tyler

ASYNCHRONY OF HATCHING IN RED-WINGED BLACKBIRDS AND SURVIVAL OF LATE AND EARLY

HATCHING BIRDS Charles Strehl

WEATHER RELATED MORTALITY OF BLACKBIRDS AND STARLINGS IN A KENTUCKY ROOSTING

CONGREGATION Baul A. Stewart

AN OBSERVATION OF POLYGYNY IN THE COMMON YELLOWTHROAT

George I . N. Rowell and //. Lee Jones

ORNITHOLOGICAL LITERATURE

president’s page

editorial: changing of the guard

ORNITHOL(JGICAL NEWS :

SERIAL PUBLICATIONS CURRENTLY RECEIVED BY THE JOSSELYN VAN TYNE MEMORIAL LIBRARY
INDEX

649

650

652

653

655

656

658

669

586

670

673

679

The Wilson Bulletin

PUBLISHED BY THE WILSON ORNITHOLOGICAL SOCIETY
VOL. 90, NO. 1 MARCH 197J^

Library

MAYS 1973

harvard

JThivbcr®'^^

The Wilson Ornithological Society
Founded December 3, 1888

Named after ALEXANDER WILSON, the first American Ornithologist.

President— Douglas A. James, Department of Zoology, University of Arkansas, Fayetteville,
Arkansas 72703.

First Vice-President — George A. Hall, Department of Chemistry, West Virginia Univer-
sity, Morgantown, W’. Va. 26506.

Second Vice-President — Abbot S. Gaunt, Department of Zoology, Ohio State University,
Columbus, Ohio 43210.

Editor — Jerome A. Jackson, Department of Zoology, P.O. Drawer Z, Mississippi State Uni-
versity, Mississippi State, Mississippi 39762. (See Ornithological News, p. 158).

Secretary — James Tate, Jr., P.O. Box 2043, Denver, Colorado 80201.

Treasurer — Ernest E. Hoover, 1044 Webster St., N.W., Grand Rapids, Michigan 49504.

Elected Council Members — Sidney A. Gauthreaux, Jr. (term expires 1978) ; James R. Karr
(term expires 1979) ; Clait E. Braun (term expires 1980).

Membership dues per calendar year are: Active, 110.00; Sustaining, $15.00;

Life memberships, $200 (payable in four installments).

The Wilson Bulletin is sent to all members not in arrears for dues.

The Josselyn Van Tyne Memorial Library
The Josselyn Van Tyne Memorial Library of the Wilson Ornithological Society, housed
in the University of Michigan Museum of Zoology, was established in concurrence with
the University of Michigan in 1930. Until 1947 the Library was maintained entirely
by gifts and bequests of books, reprints, and ornithological magazines from members
and friends of the Society. Now two members have generously established a fund for
the purchase of new books; members and friends are invited to maintain the fund by
regular contribution, thus making available to all Society members the more important
new books on ornithology and related subjects. The fund will be administered by the
Library Committee, which will be happy to receive suggestions on the choice of new books
to be added to the Library. William A. Lunk, University Museums, University of Michi-
gan, is Chairman of the Committee. The Library currently receives 104 periodicals as gifts
and in exchange for The Wilson Bulletin. With the usual exception of rare books, any
item in the Library may be borrowed by members of the Society and will be sent prepaid
(by the University of Michigan) to any address in the United States, its possessions, or
Canada. Return postage is paid by the borrower. Inquiries and requests by borrowers,
as well as gifts of books, pamphlets, reprints, and magazines, should be addressed to
“The Josselyn Van Tyne Memorial Library, University of Michigan Museum of Zoology,
Ann Arbor, Michigan.” Contributions to the New Book Fund should be sent to the
Treasurer (small sums in stamps are acceptable). A complete index of the Library’s
holdings was printed in the September 1952 issue of The Wilson Bulletin and newly
acquired books are listed periodically.

The Wilson Bulletin

The official organ of the Wilson Ornithological Society, published quarterly, in March, June, September,
and December. The subscription price, both in the United States and elsewhere, is $15.00 per year. Single
copies, SLOO. Subscriptions, changes of address and claims for undelivered copies should be sent to the
Treasurer. Most back issues of the Bulletin are available and may be ordered from the Treasurer. Special
prices will be quoted for quantity orders.

All articles and communications for publications, books and publications for reviews should be addressed to
the Editor. Exchanges should be addressed to The Josselyn Van Tyne Memorial Library, Museum of Zoology,
Ann .\rbor, Michigan. Known office of publication : Department of Zoology, Mississippi State University,

Mississippi State, Mississippi 37962.

Second class postage paid at Mississippi State, Mississippi and at additional mailing office.

Allen Press, Inc., Lawrence, Kansas 66044

Growth stages of Florida Scrub Jays. Top left; a pipped egg and two
young, age day 0. Top right: three young, age day 5, in cooling stance
with necks stretched over nest rim. Bottom left: usual banding age, day
11, when primaries are breaking from their sheaths. Bottom right: a
recent fledgling, age day 18.

THE WILSON BULLETIN

A QUARTERLY MAZAGINE OF ORNITHOLOGY
Published by the Wilson Ornithological Society

VoL. 90, No. 1 March 1978 Pages 1-158

GROWTH AND SURVIVAL OF YOUNG
FLORIDA SCRUB JAYS

Glen E. Woolfenden

Few studies of growth in passerines pertain to jays (Corvidae: Garrulinae),
and fewer still to group breeders. As a possible contribution to both topics,
I measured growth rates of young Florida Scrub Jays {Aphelocoma c. coeru-
lescens) raised by breeding groups of varying sizes. A notable exception to
the paucity of information on corvid growth is the recent work on Pinon
Jays [Gyninorhinus cyanocephalus) , a highly social, colonial breeder that
nests early in a north temperate climate (Bateman and Baida 1973). Certain
comparisons are made with this species.

Florida Scrub Jays almost invariably remain in their natal territory for
more than a year; therefore it was possible to obtain numerous post-fledging
measurements and to monitor survival of known-age jays. The growth measure
used in the analyses of survival is weight.

Florida Scrub Jays breed either as unassisted monogamous pairs or in
groups consisting of a pair and several helpers. Brood size varies from 1 to 5;
usually it is 3 or 4, and the number of feeders has varied from 2 to 8. Pairs
with helpers fledge more young than do the unassisted pairs (Woolfenden
1975).

Preceding the weight data are brief descriptions of general development
and the linear growth of certain extremities (see also frontispiece). This in-
formation should be useful for aging nestlings whose hatching dates are un-
known.

MATERIALS AND METHODS

The data on growth and survival were obtained at the Archbold Biological Station in
Highlands County, Florida, where a marked population of Florida Scrub Jays has been
under observation since 1969. Individual jays were measured daily or, in a few cases,
every other day throughout the nestling stage in 1973 when 47 nests were found, which
represented virtually all nesting attempts by 28 pairs, 25 of which produced young. The
approximately 136 eggs laid yielded 73 nestlings and 41 fledglings (1.5 per pair), almost

I IIK W ILSON lUILLKTiN • VoL 90, No. I, March 1978

all of which were measured in the course of this study. The productivity data show that
197d was a near-average breeding year ( Woolfenden 1973). Some measurements of
nestlings and fledglings also were taken in 1974 and 1975; however, except where other-
wise stated, my various analyses of growth incorporate only the 1973 measurements.

Most of the marked birds lived in open habitat, namely sparse oak scrub (Woolfenden
1969 census no. 52, 1973 >, and some jjarer.t and helper jays scolded loudly at human
intruders. Thus nest visits by investigators were brief, lasting only a few minutes, in an
attempt not to alter normal nest predation. When measuring, we removed all young from
the nest simultaneously and moved several im'ters away; this reduced the intensity of
scolding by the older jays. In addition, most nest visits were made in early afternoon
(12:30 15:30) when diurnal j)redators seemed less active, and so that several hours of
sunlight remained for the odors we left to dissipate before nocturnal predators became
active. These i)rocedures precluded obtaining detailed notes on morphological and
behavioral ontogeny.

Day 0 designates the day of hatching. Fledgling and yearling are defined as before
(Woolfenden 1973): fledgling is applied to a jay from the instant it leaves the nest
until it is 1 year old. A yearling is a jay in its second year of existence.

The ratio between the number of jays supplying food and the number of nestlings being
fed is termed the feeder index. It has ranged from 0.4 to 3.0. Nestlings wdth a 0.4 feeder
index were members of a brood of 5 fed by a pair with no helpers; the 3.0 feeder index
represents a lone nestling fed by a ])air with 1 helper.

Five measurements were taken to the nearest 0.5 mm on each young jay: length of
beak, tarsus, primary 7, and longest central rectrix (hereinafter referred to as a deck),
and w'eight. The beak was measured from the anterior end of the nostril to the tip, the
tarsus in a standard fashion as the diagonal from the joint with the tibiotarsus behind
to the joint with the middle toe in front. The primary and the longest deck w'ere measured
from the place of attachment with the skin to the tip of the papilla or feather. Falconers
use the term deck for a central rectrix and it is used here not only for brevity but also
to emphasize a function of the central rectrices of shielding the lateral rectrices from
abrasive wear. In Florida Scrub Jays the decks often become extremely ragged prior to
replacement. Weights w'ere taken with Pesola spring balances which were checked
regularly for accuracy.

The expressions significant and highly significant are used only in the statistical sense
to signify probability at the 5% and 1% levels, respectively.

Asynchronous hatching confounds the problem of measuring growth during brief once-
a-day visits to nests, and in the Florida Scrub Jay hatching of eggs from the same clutch
sometimes spans more than 1 day. In 1973 eggs from 9 of 25 nests probably had a time
span for hatching of between 1 and 2 days. For 4 of these 9 nests, a time span between
1 and 2 days was known, and for 1 additional nest a span of 2-4 days occurred between
tbe hatching cf the first egg and last egg. The nestlings were not marked until over
1 week old, and 1 assumed the smaller young in such nests were the younger. In certain
instances individual peculiarities allowed identification of these individuals.

Variation in time of fledging is another problem that vexes those who study growth
of young birds. If undisturbed, young Florida Scrub Jays remain in the nest several days
past earliest possible fledging age. In 1973, when nestlings were handled daily or every
other day, almost all fledged when 15 cr 16 days old, and only 1 healthy nestling remained
through day 17. In 1975, when young jays were handled only on day 11, and thereafter
nests were checked from convenient distances, most young fledged when 17 or 18 days
old, and a few remained through days 19 to 21. Enticing their young from the nest does
not seem to be an important part cf Florida Scrub Jay bebavior, and perhaps many nests

Woolfenden • FLORIDA SCRUB JAYS

are vacated between days 14 and 19 because of exogenous disturbances. Earliest fledging
has occurred between days 12 and 14 (1 brood), the latest on day 20 or 21 (2 broods).
These generalizations are based on observations at approximately 120 successful nests.

GENERAL DEVELOPMENT

At hatching Scrub Jays are naked, as typifies jaybirds, and the skin is
reddish-pink, nearly identical in color to the skin of a person’s hand when
flushed with blood. The viscera, especially the liver, show clearly through
the thin and weakly pigmented skin. The beak and legs are pale yellow.
Through day 1 they get darker pink, then during days 3^ yellowish pink.
On day 2 they match the color of one’s palm when it is drained of blood,
and on day 3 they have the color of jaundiced human skin. During day 3 the
skin becomes purple-black, usually on the back first, and the young match in
color heavily bruised human skin. Darkening progresses through days 4-6
until the young are dark gray-black over most of the body, and especially
dorsally. The beak becomes shiny black, the podothecae dull black.

Parting of the eyelids in nestling Scrub Jays is gradual and variable. For
a few individuals the process begins as early as day 2, for some the eyes
are still closed on day 9. For the majority the eyes open during days 4
through 7. Often in the same individual one eye begins opening before the
other.

GROWTH

Tables 1 and 2 and Fig. 1 summarize data on growth of nestlings; they
are based only on young hatched in 1973. Table 3 presents information on
fledglings, for which birds hatched both in 1973 and 1974 were used. In
order to provide a smooth transition in the growth data from nestlings to
fledglings, the 0.5-month-old young in Table 3 are the same 20 15-day-old
young in Tables 1 and 2.

Feathers. — Papillae of several tracts on the dorsum protrude prominently
from the skin about day 4. Primary 7 is at least 1 mm long by day 7 ; the
longest deck is at least 2 mm long by day 13 ( Table 1). Feathers of the
femoral and posterior dorsal tracts and secondary coverts of the alar tract
usually break from their sheaths about day 9, with the earliest on record
day 7. Primary 7 erupts between days 10 and 13, usually on day 11, and
the decks erupt between days 11 and 15, usually on day 14. Jhis rate and pat-
tern of feathering means that young Florida Scrub Jays are only sparsely
feathered until hut a few days before they fledge.

The longer 7th primaries of adult-plumaged jays (age in months 24+)
versus 6-12-month-old birds (Table 3) are new feathers, the original primaries
having been replaced during the complete second prehasic molt. The juvenal

THK W ILSO.N lUlLLF/riN • VoL 90, No. J, March 1978

I’kimaky

Table 1

7 AND Dkck Lkngtiis (mm) of Nksti.inc Flokida Sckuh Jays

A«t* in clays

Seventh primary

Longest deck

S.I).

Range

S.D.

Range

0.01

0.07

0-0.5

0.45

0.52

0 1.5

1.49

0.92

0 4

3.09

1.64

1-7

5.84

1.83

4-10

9.33

2.28

5.5-15

0.02

0.16

0-1

12.96

2.55

8-18

0.33

0.57

0-2

16..59

2.41

13-22

0.94

1.36

0-4

21.06

2.59

17-28

2.77

1.59

0-5

25.38

2.42

22-31

5.28

1.77

2.5-9

30.85

2.87

26-37

8.25

2.27

5-13

34.05

1.79

32-38

10.90

3.09

7-15

39.13

3.64

36-46

15.25

3.99

9-20

decks exist only a short time after attaining full length before they are replaced
during an incomplete first prehasic molt. The longer decks of adult-plumaged
jays versus 6-12-month-old birds are new feathers, the decks of the first basic
plumage having been replaced during the complete second prehasic molt.

Beak and foot. — The beak of Florida Scrub Jays is little more than half
full size at fledging (Table 2), and continues to grow for almost 2 months
after hatching (Table 3j. As the measurement taken includes both the integu-
mentary rhinotheca and the premaxillary bones, the increase in length shown
between ages 3-4 months and 6-12 months may merely reflect changes in
the rhinotheca. Feeding independence is a slow, gradual process in Florida
Scrub jays, which is not fully attained for about 3 months after fledging.
Perhaps this behavior reflects the slow growth of the feeding organ.

Fhe tarsometatarsus grow s rapidly ( J'ahle 2 I and attains 94% adult length
at fledging (Table 3j. Within a few days post fledging Florida Scrul) Jays
escape predators by scurrying off beneath the brush. The rapid development
of the leg probably accommodates this behavior. The measurements sum-
marized in Table 3 suggest continued slow growth of the tarsometatarsus
for many months or even a year post fledging, which, if real, probably reflects
lengthening of the bone and not changes in the integument.

1 did not measure wings of live nestlings: however, 4 specimens, age 10

Uooljenden • FLORIDA SCRUB JAYS

Beak

Table 2

AND Tarsus Lengths (mm) of Nestling Florida Scrub Jays

Age in days

Beak

Tarsus

S.D.

Range

S.D.

Range

3.00

0.10

2.5-3.5

8.81

0.35

8-9.5

3.20

0.32

3-4

9.66

0.63

8.5-11

3.59

0.33

3-4

10.97

0.90

9.5-13

3.93

0.36

3-4.5

12.73

1.06

10.5-16

4.38

0.38

4-5

14.78

1.35

13-19

4.81

0.35

4-5.5

16.71

1.18

14.5-19.5

5.22

0.44

4..5-6

19.33

1.54

17-23

5.73

0.43

5-7

21.67

1.62

19-25

6.23

0.44

5.5-7

23.89

1.36

2L.5-26.5

6.55

0.47

6-7.5

26.69

1.74

23.5-29.5

7.02

0.51

6-8

28.32

1.63

26-31

7.37

0.48

6.5-8

30.34

1.62

28-34

7.91

0.42

7-9

31.80

1.39

29-35

8.15

0.56

7-9

33.44

1.62

30-37

8.60

0.53

7-9

34.85

1.08

32-36.5

8.85

0.56

8-10

35.48

0.91

34-37.5

9.69

0.37

9-10

36.38

1.22

34-38

days, have carpometacarpi that average 71% adult length. Tarsometatarsi at
age 10 days average a similar 74% adult length. Young Scrub Jays cannot
fly for many days after leaving the nest, but this may be caused by factors
other than retarded growth of wing bones, such as slow development of muscles
and feathers.

Table 3

Age and Mean Measurements (mm) of Florida Scrub Jays

Age in
months

Beak from nostril

Tarsus

Primary' no. 7

Deck

0.5

"'*8.9 ± 0.56^

**35.5 ± 0.91

**34.1 ± 1.79

**10.9 ± 3.09

0.5-1

**9.8 ± 0.52

36.6 ± 1.21

**42.1 ± 5.60

**20.2 ± 7.92

1-2

9-13

15.3 ± 0.56

36.7 ± 1.32

**82.2 ± 1.77

*124.6 ± 6.37

3-4

27-30

**15.5 ±0.86

36.7 ± 1.34

84.8 ± 2.20

129.4 ± 4.93''

6-12

10-11

17.0 ± 1.15

37.0 ± 1.52

**83.7 ± 2.72"’

**128.9 ± 5.58"

24+

36-56

17.5 ± 0.95

37.7 ± 1.36

87.5 ± 2.86'’

134.9 ± 5.56"

S = means followed by 1 standard deviation.

Asterisks mark each mean that is significantly different from the mean immediately beneath it (f-
test ) .

\V = \vear may have caused reduced length from previous age class.

D = different feathers from tho.se measured for previous age category.

M = molt of juvenal decks begins at about age 3 months.

TllH W 1 1, SON lUil.LKTIN • VoL 90, Nu. I, March 1978

U cig/it. — 'I'lie firouth measurement analyzed in ptreatest detail is weight.
As a base for comparisons the weight of adult-plumaged jays is described
first. An adult-plumaged Florida Scrub Jay weighs 79.2 g (s.d. = 4.86j.
d'he ()0 weights chosen for this determination are of 5 live jays of each
sex for 6 bimonthly periods ( Dec.-jan., etc.). All individuals chosen
were at least 2 years old and appeared to he in good health. The
sample range (65.6-92.0 g) encomi^asses the 283 live weights available for
Florida Scrub jays 2 years old and older. Variation in weight with sex in
Scrub jays was shown by l^itelka ( 1951 I for various of the western races,
hut he had only 3 weights for the Florida race. 3 he 30 males 1 used to deter-
mine “adult” weight averaged 81.7 g (s.d. = 4.09, range 74.1-92.0), the
females 76.7 g (s.d. = 4.25, range 65.6-141.5). The weight difference between
sexes is highly significant ( t = 4.67). Seasonal variation in weight of adult-
plumaged jays, sexes combined, is graphed at bimonthly intervals (Fig. 2 ) ;
no significant differences among the 6 samples were found.

Bent ( 1916) describes the color, shape, and size of Florida Scrub Jay eggs,
hut gives no weights. The mean size of 26 eggs laid in my study tract in
1973 and 1975 (27.5 X 20.5 mm ) is similar to that for Bent’s sample of 46
eggs taken from various localities in Florida prior to 1946 (27.5 X 20.3 mm).
I he weights of 32 fresh eggs, all measured within 1 day after laying in 1973,
averaged 5.81 g ( s.d. = 0.66, range 4.1-7. 1) ; the weights of 27 eggs in the
process of hatching (not necessarily the same eggs) averaged 5.03 g (s.d. =
0.55, range 4.3-6.2 ) . For these samples weight loss from the time of laying
to the time of hatching averaged 13.3%. Eight eggs weighed when fresh and
also during hatching sustained weight losses ranging from 6.8 to 18.9%, with
a mean loss of 12.8%, which is similar to the 13.3% registered for the larger
hut less controlled sample.

According to Nice (1943:74) fresh eggs weigh 8-12% of the adult female
in passerines weighing up to 135 g. Nice deleted from her summary corvids
weighing over 135 g, which had lesser percentages (2.5-5%). For Florida
Scrub jays, fresh eggs weigh 7.6% of adult females. In the Pihon Jay fresh
eggs (x = 6.65) weigh 6.4% of adults (x = 103.3), both sexes included
(calculated from Bateman and Baida 1973). Perhaps corvid eggs tend to
constitute a smaller percentage of adult weight regardless of size. Four unfed
hatchlings averaged 1.19 g (range 3.6— 1.5 ) . The weight of additional unfed
hatchlings was estimated by sul)tracting the mean weight of moist empty
shells, taken from eggs that failed to hatch, from the mean weight of hatching
eggs. Seven fresh empty shells averaged 0.5 g (range 0.4-0.7j, and the
hatching eggs averaged 5.03 g, giving an estimated average value of 4.53 g.
At 78% the weight of a fresh egg ( 5.81 g), hatchling Florida Scrub jays are
within the range listed for certain other passerines: Lanius ludovicianus

Woolfenden • FLORIDA SCRUB JAYS

Table 4

Weights of Nestling Florida

Scrub Jays

A'^e in
days

Weight (g)

Age in days X

Weight (g)

S.D.

Range

S.D.

Range

4.82

0.73

3. 5-6.5

38.22

4.67

28.4-49.0

6.83

1.01

5.4-9.2

41.97

4.86

32.6-55.6

9.33

1.42

7.3-13.9

47.26

5.79

36.7-59.0

12.77

2.55

9.1-23.0

51.18

4.39

41.5-60.3

16.03

2.47

12.1-24.6

55.93

4.87

47.0-64.0

19.61

2.87

14.3-25.4

56.55

4.42

49.0-68.0

23.80

3.46

17.3-31.2

59.62

4.89

51.0-67.0

29.03

4.11

21.4-35.2

59.75

2.25

57.0-64.0

32.77

4.11

24.4-40.6

73-75% (Miller 1931), Molothrus ater and Quiscalus quiscula 73% I Wether-
bee and Wetherbee 1961) and Agelaius phoemceus 79% ( Holcomb and Twiest
1968). The 95% figure obtained for Piiion Jays is suspect as pointed out
by the authors. Hatchling Florida Scrub Jays weigh 6% of an adult’s weight,
which also is within the range for certain other passerines at 6-8% (Nice
1943) including the Pihon Jay (Bateman and Baida 1973).

The weight data for nestlings obtained in 1973 are summarized in Table 4
and graphed in Fig. 1. The day 0 weights used were taken almost entirely
from nestlings that had received food before weighing. These compilations
obscure the considerable variation that exists in the number of helpers and
nestlings that a given pair may have. These important variables are discussed
below.

Ricklefs (1968) found that growth for 2 corvid species was best described
by the logistic equation:

2 -|- 0~K(t\V — to)

where W is the weight of the bird in grams at the age t„ ( in days ) , A is the
asymptote of weight (g) approached by nestlings, e is the base of natural
logarithms, K is a constant proportional to the specific rate of growth, and
to is the age in days at the point of inflection on the growth curve. The
procedures outlined by Ricklefs (1967) were used by Bateman and Baida
(1973) and in this study with similar results. For the Florida Scrub Jay,
A is 60.0 (78.9 for the Pihon Jay). The age at which half of A is attained
(to) is 8.2 days (7.6 for the Pihon Jay). The overall growth rate index (K)
for the Florida Scrub Jay (0.335) is similar to that for the Pihon Jay (0.328)

1 IIK WILSON HliLLKTIN • Vol. <)0, Nu. I, March I97H

Fig. 1. Weights of nestling Florida Seruh Jays. In the diagram the single vertical line
represents the range of observations, the cross line the mean, the open column 1 standard
deviation, and the figure atop each vertical line the sample size. Below each mean, starting
with day 1, are 2 points which represent the weights of 2 starving siblings.

ami the magitie Pica pica (0.332), but larger than that for the crow Corvus
brachyrhynchos (0.172), a large, slow growing passerine.

An inverse measure of the overall rate of nestling growth (K) is the time
retiuired to grow from 10% to 90% (tio-.Ki) of the asymptote ( Ricklefs
1968 1 . Based on Bicklefs’ regression etiualion of ti(»-<M» on body size, the
crow grows more slowly than expected ( observed = 25.5 days, expected = 21.5
days), the magpie more rapidly (observed = 13.3, expected = 17.7), the
Pinon Jay essentially as expected (observed = 13.4, expected = 13.3), and
the Scrub Jay slightly slower ( observed = 13.1, expected^ 12.3).

Woolfenden • FLORIDA SCRUB JAYS

Fig. 2. Annual fluctuation in weights of “adult” and fledgling Florida Scrub Jays. The
lines connect the bimonthly means for “adults” (upper) and fledglings (lower). The
vertical bars represent 95% confidence levels. The figures denote the bimonthly sample
sizes for adults (above the lines) and fledglings (below the lines). Fledglings were
weighed from day of departure from the nest (left side April-May sample) through 1 year.

The ratio between the asymptote and adult weight describes develoimient
at fledging. The Florida Scrub Jay at 0.76 is similar to the Pihon Jay at
0.79 (Bateman and Baida 1973), and below the values obtained for 42 of 56
other passerines ( Ricklefs 1968, Table 2, R value). Low values correlate with
adult foraging and fledgling escape tactics, namely moving about on the
ground in search of food and eluding predators by running. An additional
factor may be the location of the nest, with early fledging of species whose

rilK W IKSOiN HULLK'riN • VoL 90, No. I, March 1978

Wkk.iits \M) tiik

J'amle 5

Fkkdkij Indkx ioh Nks

TI.INC

; Flohida

.ScKUH Jays

.\«c in (

cler index 0.5- 1.4
Weight (g)

Feeder index 1.. 5-3.0
Weight (g)

lays N

S.D.

4.74

0.66

5.09

0.91

6.78

0.98

7.13

1.21

9.34

1.48

9.34

1.18

12.62

2.67

13.63

1.62

15.72

2.43

17.51

2.21

18.90

2.84

21.39

2.59

23.16

3.15

26.32

3.62

28..58

3.96

31.30

4.42

32.24

3.90

35.70

4.33

37.51

4.43

41.90

4.47

41.02

4.31

47.35

4.51

45.98

5.21

52.83

5.16

50.24

4.15

55.42

2.76

54.60

4.66

60.35

2.31

55. 1 5

3.62

61.20

3.72

58.29

4.78

62.72

3.88

* and ** indicate significant differences at the .05 and .01% level, respectively.

nests are more accessible to predators. Scrub Jays do not achieve adult weight
for many months post fledging (Fig. 2).

Weights of fledgling and adult-plumaged jays are plotted at bimonthly
intervals for 1 year (Fig. 2). The fledgling weights include only those of
jays up through 1 year of age from the 1973 and 1974 year classes. The
sex of many of these fledglings was unknown, however at age 1 year the sex
ratio of Florida Scrub Jays seems to he equal (W’oolfenden 1975), and there-
fore all available weights were used.

Covariance analyses (a = 0.05 I of the data graphed in Fig. 2 reveal the
following relationships. Weights of “adults” from May through August-
Seiftemher are statistically indistinguishable from the weights of “adults”
taken from Octoher-Xovemher through the following April-May; thus all
“adults” are treated as 1 unit in the comparisons with fledglings. The weights
of fledglings taken from time of fledging in May through August-Septemher
are neither coincident nor parallel with the weights of fledglings taken from
Octoher-Xovemher through the following April-May, and the same is true of
their relationship to the weights of all “adults.” The weights of fledglings
taken from Octoher-Xovemher through the following .\pril-May also are
non-coincidenl fp < 0.05 1 with the weights of all “adults,” hut they are

Wooljenden • FLORIDA SCRUB JAYS

Table 6

Weights and the Feeder Index for 130 Day 11 Nestling Florida Scrub Jays

Year

Feeder index

Weight (g)

S.D. Range

1973

0.5-1. 4

**45.98

5.21

36.7-55.0

1.5-3.0

52.83

5.16

47.1-59.0

1974

0.4-1.4

*38.95

7.13

29.6-49.2

1.5-2.7

48.10

5.05

40.6-54.5

1975

0.5-1.4

*43.49

7.44

24.6-56.2

1. 5-2.0

48.32

5.22

39.4- .58.4

Asterisks ( * ) mark each mean that is significantly different from the mean immediately beneath it
( f-test ) .

parallel. These analyses support the conclusion that after gaining rapidly
from fledging until August-September, the young jays level off at a weight
below that of “adults.” Inspection of the bimonthly samples (Fig. 2) reveals
that between October-November and the following April-May, fluctuations
in the weights of “adults” and fledglings tend to he parallel. The only bi-
monthly change that is significant is for fledglings between December-January
and February-March ( t = 2.36 ) ; however the concomitant gains and losses
by the 2 age-classes suggest that the changes may be real.

In 1973 significant differences in weight existed between chicks with low
(0.5-1.4) and high (1.5-3.01 feeder indexes almost daily from day 4 to
day 14. Significant differences apparently occur most years as evidenced by
weights of day 11 chicks for 3 consecutive years (Table 6), including 1974
when the feeder index ranged from 0.4 to 2.7 and 1975 when the range was
0.5 to 2.0. Day 11 was chosen for time of weighing because it is late enough
in the nestling cycle for differences in weight to have developed, hut early
enough that handling the young does not cause early fledging. Day 11 also
is a convenient age for banding. Few other weight data useful for comparing
years were obtained, and none was analyzed.

MORTALITY FACTORS

Though difficult to measure, starvation of nestlings seems a minor cause
of death in Florida Scrub Jays. Based on once-daily or less frequent visits
to nests during the 6 years 1970-1975, 33 of 342 nestlings ( 10%) disappeared
from broods known to have had a continued existence. Such gradual attrition
of broods probably includes almost all nestling starvation, hut also includes
deaths caused by genetic defects, diseases ( including parasitism ) , and some

rilK WILSON imi.LKTIN • Vol. W, No. 1, March 1978

pr(‘(lali<)M. riiiis starvation apparently kills consideral)ly less than 10% of
Florida Send) Jay nesllin^2:s.

As a measure of starvation relative to hrood size, gradual disappearance
of young uas measured only from nests without helpers and with l)roods of
different sizes. For unassisted j)airs with hroods of 2 (n = 11), 3 (n = 25),
and I (n = 15) nestlings, the numher of young lost from continuing l)roods
is similar at lo, 21, and 17%, respectively. As the feeder index decreases from
1.0 to 0.5 for these unassisted ])airs with 2 versus 4 nestlings, this independent
analysis suggests that food provisioning for nestlings is not a factor critical
to rei)roductive success in the Florida Scruh Jay. However, food availability
probably has selected for clutch size which averages only 3.4 ( Woolfenden
1073).

During 7 years of watching nests, 2 breeding attempts have produced
grossly underweight broods, and both of them fledged. In 1973 an unassisted
pair fledged an underweight brood of 2. The young were far below normal
weight a few days after hatching and soon appeared weak and sick. Their
weights are plotted separately in Fig. 1. Growth of extremities, as well as
weight, were retarded. At age 15 days both young were below the minimum
recorded for all 4 linear measurements taken on heavier and relatively healthy
young (Tables 1 and 2), and measured as follows; primary 7, 24 and 18.5
mm, deck, 3 and 0 mm, beak, 7.5 and 7 mm, and tarsus 32 and 29 mm, and
weight 41.5 g and 28.0 g, respectively. These young fledged during days 21
and 20, respectively, and the lighter weight individual in all probability died
within a few’ days. The heavier fledgling died at age 99 days, at which time
he weighed only 49.6 g. However on day 82 he weighed 73.1 g, which is
almost normal for that age (see Fig. 2). A heavy helminth burden may have
contributed to its death ( see Kinsella 1974, specimen GEW 4804 ) .

The male (-WWS) of this breeding pair appeared to be a poor provider
who seemed to spend an inordinate amount of time perched near his nest.
Jhree years earlier, as a semi-independent fledgling in his natal territory
human occupants of a nearby cabin provided the jays with a bountiful supply
of peanuts. At that time I noted that this bird rarely foraged for animal food
as do other young fledglings; possibly he never gained the foraging efficiency
or drive necessary for feeding young.

In 1971 a case of bigamy resulted in the fledging of a brood of 2 under-
weight and sickly young, both of which died within days of fledging. Details,
including weights of the nestlings, are given by W oolfenden (1976).

In both these instances it seems that abnormal behavior of the breeding male
resulted in failure to provide sufficient nourisbmenl to young, even though
the feeder indexes at 0.5 were not abnormally low. Under normal circum-
stances breeding female Scrub Jays spend a large percentage of their time

\f ooljemlen • FLORIDA SCRUB JAYS

at the nest (unpubl. data), especially early in the nestling cycle. Perhaps this
general tendency prevented these 2 females from leaving their nests to forage
and thereby compensate for the inadequacies of their mates. The point of
interest here is that even when breeders exhibit abnormal behavior resulting
in grossly undernourished nestlings, fledging can occur in the absence of
predation.

In his analysis of 6 passerine species, Ricklefs (1969) identified only a few
causes for nestling loss other than starvation and predation. Two of these,
desertion and weather, are easily identified for Florida Scrub jays and are
known to he rare. By elimination, predation accounts for about o()% of all
nestling losses in the population, a percentage that is considerably higher than
the 66% tallied for the other 6 species. The high rate of nestling predation
sustained by Florida Scrub Jays probably selects strongly for a breeding
regime that reduces such losses.

SURVIVAL

Previously, survival through the first year of life was compared to adult
mortality (Woolfenden 1973), based on a sample of 143 young from 4 year
classes ( 1969-1972 ) . Now, with a sample of 269 young from 6 year
classes, 1970-1975 (the 1969 sample which is small is deleted to reduce
chance of bias), and many weight data, it is possible to examine survival as
related to nestling weight, the feeder index, and the presence of helpers.

Table 7 summarizes information on survival to feeding independence, which
virtually always is accomplished by August at age three months, for 115 of
the 130 young whose weights as nestlings are shown in Table 6. No differences
in survival are evident among the various weight groups. Perhaps with very
large samples the lightest-weight fledglings could be shown to be faring less
well, and the same might be true for the heaviest young. However, neither
the Mann-Whitney U-test nor the Wilcoxon 2-sample test demonstrated sig-
nificance with the present sample.

The feeder index can be used in an indirect method of comparing weight
and survival. As shown in Tables 5 and 6, nestlings from families with a
high feeder index (1.5-3.0) weigh more than nestlings from families with
a low feeder index (0.4-1.4). Even though weights were obtained for only
a small portion of the young jays that have been banded and censused, the
feeder index is known for virtually all. Thus the sample of young whose post-
fledging survival to independence was monitored more than doubled (115 to
267 j when the feeder index was used as an indication of high or low weight.
The number of year classes available for testing also increases, from 3 (1973-
1975) to 6 (1970-1975). The data are arranged in Table 8, and again no
cause-and-effect relationship is evident; indeed survival plotted against the

'11 IK WILSON lUILLK'l'IN • Vol. 90, No. I, March 1978

1 I

Lamle 7

l’oST-M.KI)(;iN(; SlIKMVAI. (I

iK 115 Ki.okida

.ScKUij Jays Akhanckd hy

Day 11 Wkight

WViuht (U)

Total

No. of

Percent

on Day 11

llcclKlings

independent yonnK

survivinj'

Sr> 60

.SO .S t

4S 49

40 4t

.35-.S9

.30-34

25 29

feeder index yields a straight, horizontal line. These data support the premise
that weight of nestlings has little effect on their later survival. Snow (1958)
came to similar conclusions from his study of Blackbirds (Turdus nierula) .
A regression analysis between feeder index and weight was not made because
recent field work shows that merely counting the number of jays bringing
food to a nest is an oversimplification in that amount of food brought varies
with age and sex of individual jays ( Stallcup and Woolfenden in press).

Florida Scrub Jay helpers do help, and they do so by increasing the
reproductive output of breeders, usually close kin, with whom they affiliate
(Woolfenden 19751. This conclusion, based on data from 1969 through 1973,
is further supported by similar analyses of unpublished data obtained in 1974

Table 8

Post- FLEDGING Survival

OF 267 Florida Scrub Jays Arranged by

THE I'eeder Index

P'eeder index

Total

fledglings

No. of

inclependent young

Percent

surviving

3.0

2.7

2.5

100

2.3

100

2.0

1.7

1.5

1.3

1.0

0.8

0.7

0.5

W oolienden • FLORIDA SCRUB JAYS

Post-fledging Survival

Table 9

OF 269 Florida Scrub Jays from
Helpers

Families With

AND Without

Helper status

Total

Feeder index nestlings

Percent

fledglings

Percent

independent

young

No helpers

2.0

1.0

0.7

0.7-2.0

123

Helpers

2.0

1.0

0.7

0.7-2.0

146

and 1975. But the help helpers provide the young has little to do with food
needed and food supplied (Tables 7-8). To further substantiate this phe-
nomenon, I stabilized the weight variable by measuring survival of young
jays only in families with the same feeder index, some of which had helpers,
some of which did not (Table 9), and the difference between production of
fledglings by families with and without helpers is highly significant (X“ =
17.9). Survival of fledglings to feeding independence also is greater for the
young from families with helpers although the differences are significant for
only 2 of the feeder index categories (2.0 and 1.0), and not for the third
(0.7) or for all 3 combined.

As a separate analysis, loss of clutches was measured for pairs without and
with helpers with similar results: 34% of 93 nesting attempts by families
without helpers were destroyed prior to hatching, but only 23% of 120 by pairs
with helpers. The difference is significant (x“ = 5.7). As eggs do not starve,
this provides further evidence in opposition to the hypothesis that food
provided relative to food needed limits Florida Scrub Jay reproductive success.

A question that remains is: How do helpers help increase reproductive
output if it is not by means of providing the food necessary for survival of
young? In the preceding section on nestling mortality, predation was iden-
tified as the factor responsible for about 80% of all nest losses. Suspected
nest predators include Fish Crows [ Corvus ossifragus ) , possibly Blue Jays
{Cyanocitta cristata) , certain snakes and mammals, and Scrub Jays them-
selves (Woolfenden 1973, 1975). Scrub Jays have an elaborate active nest
defense that includes scolding, plumage displays, mobbing, and outright attack,
all of which suggest these jays are capable of dissuading certain nest predators.
Jhus I suggest the major way that Florida Scrub Jay helpers help is by

THE W lESON HUEEETIN • Vol. 90, Nu. I, March 1978

decreasing predation on the nests containing eggs or nestlings, and to some
extent on the scattered fledglings, of the breeders with which they affiliate.

CONCLUSIONS

(Growth of nestling Florida Scruh Jays seems typical of passerines their
size. Minor developmental features common to both Pinon and Scrub jays,
which Bateman and Baida ( 1973) consider adaptive for breeding during cold
weather by Pinon Jays, seem adaptive for breeding in a bot sunny climate
by Florida Scrub Jays. Ihus dark skin pigmentation and tbe more rapid
development of dorsal feathers than those of the venter may help shield
nestlings from harmful (luantities of ultraviolet light. At fledging Scrub Jay
young are less developed than most passerines studied thus far ( Ricklefs PJ68),
and although comparative data are few, post-fledging growth seems retarded,
as evidenced by the failure of young to attain adult weight by the end of a
year. J be social organization of the population probably allow s for gradual
growth, and indeed it may even cause it. If an advantage exists for gradual
growth, the security of a defended natal territory may permit it. As an
alternative hypothesis, intrafamilial dominance hierarchies relegate fledglings
to subordinate positions ( Woolfenden and Fitzpatrick P977), which may
suppress their gaining weight. It remains to be established whether or not
these hierarchies result in higher survival of birds of particular weights.

Most Florida Scruh Jay nests fail (Woolfenden 1973), hut rarely because
of desertion or weather. Starvation accounts for less than 10% of all
nestling losses while predation apparently accounts for over 80%. Young
fed by relatively more feeders are heavier, hut weight and the feeder index
do not affect post-fledging survival. However, survival to fledging is related
directly to the existence of helpers (Woolfenden 1975 and Table 9). Pre-
liminary observations indicate the amount of help helpers provide varies
with sex and age; therefore more refined measures of success relative to
number of helpers are omitted intentionally. It is postulated that helpers
assist breeders by reducing nest predation. The possibility that group breed-
ing results in direct advantages to tbe breeders and the helpers is currently
under investigation.

SUMMARY

Growth of young was measured in a marked population of Florida Seruh Jays that has
been censused from 19f)9 to the present. Data were gathered mostly in 1973 when samples
ranged up to 59 nestlings, whieh were the reproduetive efforts of 28 pairs.

Fresh eggs weigh 5.8 g, 7.()% of adult female weight, and lose about 13% of their weight
during incubation. Newly hatehed, unfed young weigh about 4.5 g, about 78% of a fresh
egg. Based on growth curve computations, nestling growth is half completed at 8.2 days.
Tin* overall growth rate index of 0.335 is similar to that of Pinon Jays, and the young

Wooljenden • FLORIDA SCRUB JAYS

grow only slightly slower than expected for their body size. Development at fledging lags
behind most passerines thus far measured, as is true also of Pihon Jays. Florida Scrub
Jays do not attain “adult"’ weight during the first year. Growth of certain extremities
also re(juires many months. Fluctuations in fledgling weights parallel those of “adults”
from fall to early spring.

Desertion and weather rarely cause nesting failure, and starvation of nestlings accounts
for less than 10% of nestling losses. Predation is the major factor; it accounts for over
80% of all nestling losses.

Breeding pairs with helpers produce more young, especially fledglings, than do unassisted
pairs. Nestlings fed by relatively more feeders are heavier, hut survival as fledglings does
not correlate with nestling weight or the feeder index. Even nestlings half normal weight
at day 11 appear to survive as fledglings as well as do heavier birds. Decreasing predation,
especially on nest contents, is proposed as the major way that helpers increase reproduction.
Elaborate active nest defense by breeders and helpers supports the suggestion. The sus-
pected predators they may sometimes dissuade are certain snakes. Fish Crows, Blue Jays
and Scrub Jay cannibals.

ACKNOWLEDGMENTS

As is true for the earlier work, this phase of a long-term life histor> study was completed
through the generosity and interest of Richard Archbold, Resident Director, and James
N. Layne, Director of Research, of the Archbold Biological Station. Release time from
teaching during Spring quarter 1973 was made possible through a Research Council Award
of the University of South Florida. Additional support came from the Frank M. Chapman
Memorial Fund and the St. Petersburg Audubon Society.

Susan C. White and Stephen A. Bloom provided invaluable advice on mathematical
procedures and D. Bruce Barbour, Anthony R. DeGange, John W'. Fitzpatrick, Jerre A.
Stallcup and Chet E. Wdnegarner helped with the fieldwork. Ralph W. Schreiber and
Susan C. W bite improved the manuscript. I thank all of these persons and institutions
for their help. The help of the referees, D. F. Caccamise and R. E. Ricklefs, is gratefully
acknowledged.

LITERATURE CITED

Bateman, G. C. and R. P. Balda. 1973. Growth, development, and food habits of young
Pihon Jays. Auk 90:39-61.

Bent, A. C. 1946. Life histories of North American jays, crows, and titmice. U.S. Natl.
Mus. Bull. 191.

Holcomb, L. C. and G. Twiest. 1968. Red-winged Blackbird nestling growth compared
to adult size and differential development of structures. Ohio J. Sci. 68:277-284.

Kinsella, J. M. 1974. Helminth fauna of the Florida Scrub Jay: host and ecological
relationships. Proc. Helminthol. Soc. Wash. 41:127-130.

Miller, A. H. 1931. Systematic revision and natural history of the Ameriean shrikes
^L(iniiis). Univ. Calif. Publ. Zool. 38:11-242.

Nice, M. M. 1943. Studies in the life history of the Song Sparrow. 2. The behavior of
the Song Sparrow and other passerines. Trans. Linn. Soc. N.Y. No. 4.

PiTELKA, F. A. 1951. Speciation and ecological distribution in American jays of the
genus Aphelocoma. Univ. (ialif. Publ. Zool. 50:195-464.

Ricklefs, R. E. 1957. A graphical method of fitting eipiations to growth curves. Ecology
48:978-983.

l‘» 'I'HK W ILSON BllLLE'riN • Vol. 90, No. 1, March 1978

. 1%8. I’attcrns of p;rowtli in birds. Ibis 110:419-451.

. 1969. An analysis of nestirifi mortality in birds. Smitbson. Contrib. to Zool.

No. 9.

.'^Now, 1), \V. 19.58. The l)rcedinfr of tbe Blackbird Tardus rnerula at Oxford. Ibis 100:

1-30.

.Stai.lci I*, J. A. AND O. E. WOoi.FKNDKN. In press. Eamily status and contributions to
breedings by Elorida .Scrub Jays. Anim. Behav.

W4:tiikhhkk, I). K. and N. S. Wetiikkuke. 1961. Artificial incul)ation of eggs of various
bird species and some attributes of neonates. Bird-Banding 32:141-159.

Wooi.FENDEN, (i. E. 1969. Breeding-bird censuses of five habitats at Archbold Biological
Station. Audubon Eield Notes 23:732-738,

. 1973. Nesting and survival in a population of Elorida Scrub Jays. Living Bird

12:2.5-49.

— 1975. Elorida Scrub Jay helpers at the nest. Auk 92:1-15,

— 1976. A case of bigamy in tbe Florida .Scrub Jay. Auk 93:443-450.

AND J, W. Fitzpatrick. 1977. Dominance in the Florida Scrub Jay. Condor 79:

1-12.

DKPT. OF BIOLOGY, UMV. OF SOUTH FLORIDA, TAMPA 33620; RESEARCH ASSOCIATE,
ARCHBOLI) BIOLOGICAL STATION, THE AMERICAN MUSEUM OF NATURAL
HISTORY. ACCEPTED 13 .JULY 1976.

REQUEST EOR ASSISTANCE

Vulture sightings. — Sightings of and information about Turkey Vultures tagged with
blue or orange streamers, each with a white letter and a one or two digit number, would be
appreciated. Tbe tags are about 3" X 6" and are fastened to the patagium with a num-
bered cattle ear tag. Birds are tagged on either the right or left wing. The tags are on
both tbe dorsal and ventral surfaces of tbe wing. Data requested include: tag number,
tag on left or right wing, date, time and place of sighting, activity of the bird and its
proximity to other birds. I am particularly interested in tagged birds seen mating or
in tbe nest. An opportunity to tag nestlings of tagged birds would be invaluable. Please
send sighting data to: Bird Banding Laboratory, Office of Migratory Bird Management,
Fish and Wildlife Service, Laurel, Ml) 20811 and/or Sheila Parness Gaby, 6832 S. W.
68 St., .S. Miami, FL .33143.

VERTICAL DISTRIBUTION OF BIRDS IN A LOUISIANA
BOTTOMLAND HARDWOOD FOREST

James G. Dickson and Robert E. Noble

Resources used by avian species are probably in limited supply in forest
ecosystems resulting in interspecific competition, in resource partitioning,
and in the segregation of species on habitat gradients ( Koplin and Hoffman
1968, Cody 1974, Schoener 1974 j. Resource partitioning has been accom-
plished through various “coexistence mechanisms” (Cody 1974). Schoener
( 1974 ) hypothesized that habitat dimensions are important more often than
food-type dimensions which are more important than temporal dimensions in
resource partitioning. One of these mechanisms or dimensions is a spatial
segregation of birds into vertical strata. Vertical height distribution is one
dimension of niche definition. MacArthur and MacArthur (1961) correlated
vegetative height diversity and bird species diversity showing how bird com-
munities responded to vegetative profiles. Tramer (1969) also noted the
response of bird populations to vegetative layering. Cody ( 1968) observed
vertical feeding height selectivity in tall vegetation, and Pearson (1971) and
Karr ( 1971 ) documented vertical stratification in tropical birds.

How does maturity of an ecosystem affect bird populations and resource
use? Odum (1969) predicted an increase in potential niches and interspecific
competition resulting from increased biomass stratification with the advance
of vegetative succession.

Another question that remains essentially unanswered is what are the sea-
sonal patterns of resource partitioning in communities. Bird energy budgets
fluctuate seasonally (King 1974) as do behavior patterns. Deciduous forests
present seasonally changing vegetative profiles and habitat structures. These
seasonal phenomena result in changes in bird spatial distributions.

The objective of our investigation was to ascertain vertical distributions of
certain avian species and to analyze the seasonal changes in these distributions
in a mature Louisiana bottomland hardwood forest.

STUDY AREA AND METHODS

This investigation was conducted on the Thistlethwaite Wildlife Management Area
between Washington and Laheau, St. Landry Parish, Louisiana. The area is an old
floodplain of the Mississippi and Red rivers. It is described as a south central Louisiana
mature bottomland hardwood forest, and classified as hardwood bottom (Braun 1950:293).
Vegetation on the area was measured on variable radii plots. The mature bottomland
hardwood forest was fully stocked (28.2 m“ basal area/ ha). Oaks iQuercus spp. ) were
dominant overstory vegetation. Cane i Arundinaria gigantea) , palmetto (Sah(d minor),
and ironwood iCarpinus caroliniuna) were primary understory species.

TllK W ILSON 1UILLP:TI.\ • Vol. 90, No. I, March 1978

Wrtical data from 4103 sightings of 26 species of birds were analyzed. Height

data were collected approximately 5 mornings per month from .lanuary 1972 to Fehruarx
1974 while c(‘nsusing birds from a 1.6 km transect. .Sightings were made from sunrise to
4 h after sunrise; therefore, no data on daily patterns in heights were gathered. Vertical
strata categories (MacArthur and MacArthur 1961) were: ground-0.6 m, 0.6-7. 6 m,

7.6 m-canopy top ( apj)roximately 2.3.9 m), and above-canopy. These zones probably corre-
sponded, as well as any, to the vegetation profile, although no distinct layers of vegeta-
tion were observable. No corrections were applied to compensate for differences in sight-
ing distances in foliage profile throughout the year, although there were decided seasonal
changes. In summer the vegetation appeared to be almost ecjually distributed at differ-
ent heights. In winter after the deciduous leaves had fallen, the ground and mid-story
vegetation, mainly palmetto and cane (both evergreen), was denser than the mostly leaf-
less canopy.

.Ringing birds were omitted in this study. Sightings were not restricted to any particular
behavioral category, although most birds were foraging when detected. There may have
been some height differences corresponding to different bird behavior, but we did not
attempt to distinguish behavior when recording heights. A behavioral division of height
classes would have reduced our sample sizes significantly. We do not believe this lump-
ing significantly affected results.

Birds were categorized into 1 of the 4 strata at the time of initial sighting with a few'
minor exceptions (Table 1). Ground occupants were often first seen in mid-air after
having been flushed from the ground. These instances of flushing were regarded as
ground sightings.

Height diversities were calculated from the information theory of Shannon (1948).
Using this formula, dispersal among classes, or diversity, was calculated, based on equal-
ity of distribution of observations among the 3 classes (above-canopy stratum excluded).
Height diversity = -^ Pi loge Pi, where Pt rz proportion of observations in the fth cate-
gor>-. For the 3 height categories used, 1.099 would represent maximum diversity or
e(jual dispersal among all categories, and conversely, a complete distribution in only 1
eategory would have zero diversity.

Birds in the “above-canopy” stratum were divided into 2 groups: those carrying on
their “normal” activities at that height and those merely relocating themselves. We in-
cluded the above-canopy stratum for Black Vultures and Common Crows (Table 1) be-
cause they appeared to regularly use that height while carrying on their “normal” ac-
tivities. Those relocating themselves in the “above-canopy” stratum were omitted from
further consideration.

Bird vertical stratifications were compared on a seasonal, species, and family basis.
The 3 strata comparisons within the forest were used for all species except the Black
Vulture and Common Crow. Comparisons were tested by means of the chi square test for
independent samples at the .01 level of significance unless otherwise specified. There
were 2 degrees of freedom in each chi srjuare test of 3 vertical strata. The basic assump-
tion of this test is that all observations w'ere independent of all other observations. We
felt that data on Common Crackles and Cedar Waxwings did not meet the basic
assumption, due to their occurrence in flocks and to our influence on their vertical dis-
tribution. As a result, they were excluded from further consideration. Flocking was ob-
served to a lesser degree in other bird groups but not to the extent to invalidate the as-
sumption of independence. The criterion for sufficient samples for reliability was taken
from .'^iegel (1936). In conq)aring the 3 vertical strata within the forest, no expected

Dickson and Noble • VERTICAL DISTRIBUTION OF BIRDS

Vertical Stratal Index of

Table 1

Common Thistlethwaite Birds Based on
Occurrence in 3 Strata^

Frequency of

Common name

Scientific name

Index-

Black Vulture

( Coragyps atratus )

3.50

Common Crow

iCorvus brachyrhynchos)

3.04

Red-headed Woodpecker

( Melanerpes erythrocephuhis )

2.80

Blue Jay

{ Cyanocitta cristata)

2.75

Fileated Woodpecker

{ Dryocopus pileatus )

2.74

Carolina Chickadee

iParus carolinensis)

2.67

Red-hellied Woodpecker

y Melanerpes caroliniis)

2.66

Tufted Titmouse

{ Paras bicolor )

2.51

Yellow-bellied Sapsucker

iSph y ra p icus va ri us )

2.39

Yellow-rumped Warbler

(Dendroica coronata)

2.34

Brown-headed Cowbird

( Molothrus ater)

2.30

Hooded Warbler

(Wilsonia citrina)

2.24

Mockingbird

( Mi m us pol ygl ottos)

2.24

Common Flicker

( Col apt es aurctus )

2.23

Ruby-crowned Kinglet

{ Regains calendula)

2.14

White-eyed Vireo

( Vireo griseus)

2.13

American Goldfinch

iSpinus tristis)

2.10

Brown Thrasher

(Toxostoma rufum)

2.08

Carolina Wren

( Thryothorus ludovicianus )

2.01

Cardinal

( Cardinalis cardinalis )

2.00

American Robin

( T Urdus migratorius )

1.94

Kentucky Warbler

(Oporornis jormosus)

1.88

Hermit Thrush

( Catharus guttatus )

1.77

Rusty Blackbird

(Euphagus carolinus)

1.74

Rufous-sided Towhee

( Pi pilo eryth rophthal m us )

1.41

hite-throated Sparrow

iZonotrichia albicollis)

1.27

V 58.88

X 2.26

^Fourth stratum (above canopy) used only for Black Vulture and Common Crow.

“ Index was computed by multiplying number of sightings in each stratum by: 1 for stratum 1

(ground— 0.6 m), 2 for strabim 2 (0.6— 7.6 m), and 3 for stratum 3 (7.6 m— canopy top).
The sum of these products for each species was then divided by total sightings, giving relative
mean height.

values of less than 1 were tolerable, ami no more than 20% of the expected values could
be less than 5. In the few instances of a low value in 1 stratum, strata were comliined
for purposes of comparison.

For comparative purposes, the strata were assigned the following values: ground, 1;

mid-story, 2; canopy, 3; and aliove-canopy, 4. Stratal index was calculated by multiply-
ing these values in each stratum by tlie frequency in each stratum. The sum of these
products divided by total frequency defined stratal index, based on frequency of ob-
servations in each of the strata.

'I'llK WII.SO.N lU'LLKTI.N • Vol. <)0, No. I, March 1978

Fa RLE 2

VKimcAi. IIkk.iit

DiVKK.sTIY of (ioMMON BiHDS IN .‘3 HeIGIIT

Catkgokiks

.Species

Diversity^

Species

Diversity

Muximum Diversity"

1.099

\ ellow-hellied Sapsucker

.702

Ameriean Kohin

1.075

Red-bellied Woodpecker

.082

(iommon Flicker

1.050

Ruby-crowned Kinglet

.078

Husty 8luekl)inl

1.051

White-throated Sparrow

.0.50

Ameriean (iuldfineli

1.014

(Carolina (diickadee

.0.34

Hrown Tliraslier

1.000

Pileated Woodpecker

.022

Hermit Tlirusli

.980

Kentucky Warbler

.010

(Cardinal

.9.50

Blue Jay

.004

Hrown lu'aded Cowl)ird

.924

Carolina Wren

.571

Moekinghird

.898

Hooded Warl)ler

.551

^ ellow-rumped Warhler

.808

White-eyed Vireo

.534

'I'ufted Titmouse

.801

Red-headed Woodpecker

.513

Kufous-sided Towliee

.792

V 18.832

X .785

1 Computed by infoniiation theory (height diversity = — ^ P^ log^ P^, where Pj = proportion of
observations in the /tli height category).

- Equal distribution in all height categories, height diversity = 1.099; distribution in only 1

height category, height diversity = 0.

RESULTS .\M) DISCUSSION

S])ecies and jam Hies. — Black Vultures had the highest mean vertical dis-
trihution ( Table 1 I . Over % of the sightings were of soaring birds above
the canopy.

Woodpeckers were predominately canopy dwellers, hut different niche
breadths in vertical distributions were evident. Of all sightings, 68% were
above 7.6 m and less than 3% were found near the ground ( Fig. 1 ). Pileated
and Red-hellied woodpeckers were similar in vertical distribution to the ag-
gregate of woodpeckers. Distributions of 3 species differed from the 2 above
species. Common Flickers were more ground oriented (21% of sightings),
and were exceeded in vertical dispersion (diversity index (1)1) = 1.056, Table
2 ) by only one bird. \ ellow-hellied Sapsuckers were primarily mid-story oc-
cupants (58%), and secondarily canopy occupants (41% of sightings) during
their winter j)resence ( 1)1 = 0.762). Red-headed W oodpeckers were the most
canopy dependent Picidae (82%) with the most restrictive vertical height
dimension of niche breadth of all birds ( 1)1 = 0.513. Table 2 ).

I'he corvids were located high in the Thistlethw aite woods (Table 1 ). Blue
Jays were closely associated uith the canopy level (stratal index = 2.75, 1)1 =

Dickson and Noble • VERTICAL DISTRIBUTION OF BIRDS

0.604). They were even more strongly canopy oriented than were the Picidae
(^“ — 12.6, P < .01). Common Crows were located even higher; 40% above
the tree tops.

The similarly distributed ( P > .01 ) Tufted Titmouse and Carolina Chick-
adee were common canopy occupants. Fifty-four % of the Tufted Titmice
and 67% of the Carolina Chickadees were observed in the canopy. They were
less frequently observed in the mid-story ( 43% titmice, 33% chickadees ) .
The Paridae, along with the Picidae, were the least frequent ground level oc-
cupants ( titmice 3% ; chickadees, none ) .

We sighted 83% of the Carolina Wrens in the mid-story. The chi scjuare
value for the comparison of wrens to the aggregate of all birds ( which was
also most numerous in mid-story sightings) was 128.6 I P < .01 ) . Supporting
this idea of mid-story association is the low height diversity of .571.

The Mimidae, Northern Mockingbirds and Brown Thrashers, were a ver-
tically diverse group ( D1 = 0.898, Northern Mockingbirds; DI = 1.006, Brown
Thrashers), tending toward the mid-story level. Over 50% of the sightings of
each were in this level.

The turdidae exhibited an unusual pattern of height distributions. Al-
though we assumed that intrusion into the woods had no influence on bird
heights, we may have had some effect on the heights at which Hermit Thrushes
and American Robins were observed. Half of the Hermit Thrush sightings
were in the mid-story and over % on the ground ( Fig. 1 ) . Perhaps a small
portion of the mid-story sightings were of birds that flew there from the
ground after being flushed. The robin was the most uniformly distributed
bird in the 3 strata ( DI = 1.075) : ground (41%), canopy (35%), and mid-
story ( 24%) .

Ruby-crowned Kinglets were common mid-story winter residents. Of 73
sight tallies, 76% were within the 0.6-7.6 m level. The ground stratum was of
little importance ( 5% ) and the canopy stratum was of medium importance
(19%). Their dispersal among the 3 strata was 0.678, near the mean of all
species (0.785).

White-eyed Vireos were the only breeding vireo commonly seen. These
birds were closely associated with the mid-story. This is shown by the pro-
portion of mid-story sightings (82%) and the low height diversity (0.534,
second lowest of all birds). Although there were insufficient sightings of
Red-eyed and Yellow-throated vireos ( V . olivaceus and V. flavifrons) for
valid conclusions, the few that were sighted, and those heard, showed a canopy
preference.

Yellow-rumped Warblers, 1 of 2 common winter warblers, were located
mainly in mid-story (54%) and canopy (40%). The 2 commonly seen breed-
ing season warblers, Kentucky Warbler and Hooded Warbler were selective

21.

TllK WILSON HLILLKTIN • Vol. 90, No. I, March 1978

SPECIES AND NO. OF SIGHTINGS
Black Vulture* (22)

Common Flicker (43)

Pileoted Woodpecker (51)

Red- bellied Woodpecker (99)
Red-heoded Woodpecker (373)
Yellow-bellied Sopsucker (176)
All Woodpeckers (802)

Blue Joy (184)

Common Crow (52)

Carolina Chickadee (55)

Tufted Titmouse (137)

Carolina Wren (148)

Mockingbird (33)

Brown Thrasher (154)

American Robin (229)

Hermit Thrush (30)

Ruby- crowned Kinglet (73)
White-eyed Vireo (45)

Yellow - rumped Warbler (82)
Kentucky Warbler (24)

Hooded Warbler (17)

Rusty Blackbird (39)

Brown -headed Cowbird (33)
Cardinal (390)

American Goldfinch (21)

Rufous -sided Towhee (82)
White - throated Sparrow ( II 36)
All Birds (4103)

PER CENT OF SIGHTINGS

*Above Conopy Stratum considered for Black Vultures and Common Crows only

I I 0-0.6m. I j 0.6- 7.6m. 7. 6m. - Canopy Top Above Canopy

FlC. 1. Vertical height distrihutions of common birds.

Dickson and Noble • VERTICAL DISTRIBUTION OF BIRDS

in their forest profile use. Over 75% of sightings of Kentucky and Hooded
warblers were in the 0.6-7. 6 m stratum and the height diversity of each was
less than the mean of all birds by more than 25%. They appeared to differ
in use of the ground stratum (4 of 24 sightings — Kentucky Warblers, 0 of 17
sightings- — Hooded Warblers), although there were insufficient data for a
valid statistical test.

Brown-headed Cowbirds were mainly a mid-story, and secondarily a canopy
occupant (DI = 0.924). Rusty Blackbirds were diversely distributed ( DI =
1.051, 34% higher than the mean for all birds); they were found on the
ground, mid-story, and canopy in decreasing order of occurrence.

The seed-eating fringillids, as expected, tended to be close to the ground.
The 2 species (White-throated Sparrow and Rufous-sided Towhee) found
most frequently near the ground were in this family. Over 75% of the White-
throated Sparrows and 62% of the Rufous-sided Towhees were detected within
0.6 m of the ground. Conversely, only 3% of the sparrows and 4% of the
towhees were detected in the tree canopies. The Northern Cardinal and the
less common American Goldfinch differed ( P < .01 ) from other fringillids.
Both were mainly located in mid-story (60%, cardinal; 52%, goldfinch), and
both showed high dispersal within the 3 strata ( I)I > 20% higher than the
mean of all species ) .

Most individual species were more specialized than the aggregate of all
birds. Of the different species investigated, only Brown-headed Cowbirds,
American Goldfinches, Hermit Thrushes, and Northern Mockingbirds did
not differ significantly ( P > .05 ) in height distribution from sightings of all
birds. These species were commonly found in all strata and as a result, ex-
hibited a greater than average height diversity.

Vertical resource partitioning. — Different vertical resource use strategies
were evident in birds in this mature ecosystem. Some species were specialists
in using 1 of the 3 strata; some used 2 strata exclusively, or nearly so; some
were found in all strata, but preferred 1 or 2 strata; and some species used
all strata almost equally. Specialist species ( i.e., those with lowest DI and
>78% of sightings in 1 stratum ) in the 7.6 m-canopy top stratum were
Pileated Woodpecker, Red-headed Woodpecker, and Blue Jay. Carolina
Wrens, White-eyed Vireos, and Kentucky Warblers were mid-story associated
species, and no species were predominantly ground dwellers.

Other species concentrated their activities in 2 strata I >94% of sightings ) .
Those found predominantly in the mid-story/canopy were: Red-bellied Wood-
pecker, Yellow-bellied Sapsucker, Carolina Chickadee, Tufted Titmouse,
Ruby-crowned Kinglet, and Hooded Warbler. White-throated Sparrows and
Rufous-sided Towhees were the ground/mid-story dwellers.

Northern Mockingbird, Brown Thrasher, Hermit Thrush, Yellow-rumped

TllK WILSON IUJLLP:TIN • Vol. 90, No. 1, March 1978

\\ arl)ler, Brow ii-lieaded ("owhird, and .Northern Cardinal were l)asically gen-
eralists in vertical selectivity, hut showed slight jireferences for 1 or 2 strata.

(a)ininon Flicker. American Kohin, American Goldfinch, and Rusty Black-
bird were generalists in vertical distribution, displaying optimum height dis-
persal among the 3 strata, and maximum niche breadth.

riie aggregate heights of all birds revealed a fairly uniform use of the 3
strata I Fig. 2). Each stratum was of approximate e(iual value as an avian
resource unit. Ibis substantiated the height units selected, and pointed out
the scaling differential of birds in vertical distribution (Cody 1974:70).
Although the canopy stratum represented 71% of the total forest height, only
33% of bird sighings were within this stratum. Conversely, the 0-0.6 m
ground stratum represented 2% of the total height and contained 31% of the
birds. Plant detritus, particularly mast, accumulated on the ground and this
provided direct and indirect food sources for the ground foraging birds.
Additionally, the solid substrate probably rendered the ground more acces-
sible to foraging birds.

d'he mid-story contained proportionately more birds than the canopy, but
fewer than the ground stratum. The continuous cover of evergreen cane and
palmetto of the mid-story, may have influenced vertical distribution. Flying
birds also appeared to prefer this height.

Seasonal vertical distribution. — Seasonal shifts in vertical distribution were
evident in Thistlethwaite birds. Due to the seasonal occurrence of some
species, and the small number of samples of many others when categorized
into seasons, we usually grouped individual species into higher taxa or on
a residency status basis.

dTiere was a gradual shift in distribution of birds upward in height through
the 3 strata from the winter season through spring to summer and a pro-
nounced downward movement from fall to winter. The comparison of win-
ter to summer showed decided differences. Spring brought a slight, but non-
significant (P>.()5), shift upward in height for the aggregate of all birds,
and for permanent residents when considered separately (Fig. 2l. For all
birds, ground detections fell from 38 to 36% and canopy detections rose
from 29 to 32% from winter to spring. In the permanent resident group,
ground detections fell from 27 to 20% and canopy sightings increased from
22 to 27% from winter to spring. Common fringillids ( hite-throated Spar-
rows. I^ufous-sided Towhees, and Northern Cardinals), which were, in part,
included in the 2 previous groupings, showed a significant ( P < .05 ) shift
upward in distribution from winter to spring. Ground detections fell from
70 to 65%, and canopy detections increased from 4 to 7%, as the birds re-
sponded to the seasonal spring flourish of primary production of trees and the
corresponding shift of available food. Birds were attracted to new vegetation

Dickson and Noble • VERTICAL DISTRIBUTION OF BIRDS

ALL BIRDS PERMANENT RESIDENT

NO. OF SIGHTINGS WINTER

I NOV. — 28 FEB.

^ 29 1

2298

33|

38 1

NO. OF SIGHTINGS

]50

258

902

SPRING

1 MAR.

15 MAY

l27 ■)

|53 i

36]

[20 )

SUMMER

16 MAY— 31 AUG.

C 35|

|27 )

j57|

^63|

1 8|

flo j

413

60 40

FALL

I SEPT— 31 OCT.

]27

20 0 0 20 40

3 — 7.6- Canopy Top
2 - 0.6- 7.6 m.

I - 0 -0.6 m.

Fig. 2. Seasonal vertical distribution of Thistlethwaite birds, ex{)ressed as % of siglit-
ings in 3 height categories.

growth of the deciduous forest. White-throated Sparrows and Northern
Cardinals were commonly observed feeding on new buds and samaras of
American elm {Ulmus americana) in March. This winter to spring height
distribution shift of the common fringillids was partially responsible for the

TilK W II.SON IUjLLKTI.N • VoL 00, No. J, March 1078

2l\

clumjie in lieifilit (lislrihulioii of other bird fJiroupinf^s in which the fring;illids
may liave been included ( ijermanent residents arul all birds).

dhe movement from firound to mid-story, and from mid-story to canopy,
was more pronounced from spring to summer. In comparing the spring to
summer distributions of all birds combined, a chi s(}uare value of 143.81
( P < .01 ) was noted. (Ground sightings declined to 8%, mid-story sightings
increased to 57%, and canopy sightings increased to 35%. 3 he increase in
stratal index of birds substantiated this upward movement. This was par-
tially due to a species change between seasons. 3 he departure of White-

throated Sparrows from February to April lessened the lower strata detec-
tions. But the permanent residents also exhibited a significant (;^- = 8.98,
B < .05) shift upward in response to the vegetation profile change.

A highly significant difference ( P < .01 ) was noted between winter and

summer vertical height distributions for common permanent residents “

26.11) and the aggregate of all birds ( = 178.78 ) . There w as a slight

change in height distribution from winter to spring, and a more pronounced
change from spring to summer. 3he winter to summer comparison embodied
these 2 lesser seasonal height distribution changes.

I he summer to fall comparison showed no discernible shift in vertical dis-
tribution of I histlethw aite birds. Ground detection percentages remained
virtually unchanged for all birds and permanent residents. Figure 2 reveals
a shift of about 16% of sightings of all birds from mid-story to canopy. We
believe this was misleading due to the autumn arrival of numerous Red-headed
W oodpeckers. This conspicuous canopy dweller inflated the number of canopy
detections. A check of the common permanent residents revealed no notable
change in vertical distribution from summer to fall = 0.66, P > .05).

W ith the accumulation of plant detritus, particularly mast, on the ground
in fall and early winter, the birds redistributed themselves at lower levels in
the i)i ofile. I here w as a highly significant difference ( P < .01 ) between fall
and w inter for all birds ( x~ = 156.25 ) and for permanent residents ( x" =
9.81).

Ibis seasonal height distribution change resulted in a corresponding change
in height diversity. A Least Sfiuares Analysis of Variance showed a highly
significant difference (F = 8.37; d.f. = 2, 3; P<.01) between seasons.
Height diversity approached maximum during w inter I 1.092 ) and spring
11.097). During summer and fall diversity was lower (0.890, 0.894), with
birds favoring the mid-story and canopy.

SUMMAUY

Vertical liei^ht data from tl03 sightings of 2() species of birds were analyzed in order
to better understand height segregations and resource use. Height categories used were:

Dickson and Noble • VERTICAL DISTRIBUTION OF BIRDS

ground to 0.6 m, 0.6 m to 7.6 m, and 7.6 m to canopy top (approximately 25.9 m). Bird
height distributions were compared J)\ means of the chi square test for 2 independent
samples. Height diversities were computed by the information theory. The most
ubiquitous species in height dispersion were: American Robin, Common Flicker, Rusty
Blackbird, and American Goldfinch. The species most restricted in the forest profile
and the zones they inhabited were: Red-headed Woodpecker, Pileated Woodpecker and
Blue Jay — canopy; White-eyed Vireo, Kentucky Warbler, and Carolina Wren — mid-story.
There was a gradual upward shift in distribution of all birds from winter through spring
to the summer breeding season. There was a highly significant winter to summer height
distribution change from a nearly e(|ual distrihuton at all levels in winter to a pre-
dominantly mid-stor>- and canopy distribution in summer. Corresponding with this was a
reduction in height diversty of the aggregate of all birds. These shifts were presumably
a response of the birds to the seasonal change in foliage piofile and food supi)ly of a
deciduous forest.

ACKNOWLEDGMENTS

The research was supported by the School of Forestry and Wildlife Management and
the Agricultural Experiment Station, Louisiana State University, Baton Rouge. We are
thankful to R. B. Hamilton, C. R. McLellan, Jr., and P. E. Schilling for valuable as-
sistance in the research design and analysis of data, and to M. Raymond for programming
the data for computer analysis.

We are grateful to the Louisiana Tech University School of Forestry for clerical as-
sistance, and to J. W\ Goertz, R. B. Hamilton, J. A. Jackson, 1). James, and J. R. Karr
for valuable suggestions with the manuscript,

LITERATURE CITED

Braun, E, L. 19.50. Deciduous forests of eastern Noith America. Blakiston Co., Phil-
adelphia.

Cody, M. L. 1968. On the methods of resource division in grassland bird communities.
Am. Nat. 102:107-147.

— . 1974. Competition and the structure of bird communities. Princeton Univ.

Press, Princeton, N.J.

Karr, J. R. 1971. Structure of avian communities in selected Panama and Illinois hab-
itats. Ecol. Monogr. 41:207-233,

King, J. R. 1974. Seasonal allocation of time and energy resources in birds. In Avian
energetics ( R. A. Paynter, Jr., ed.), Nuttall Ornithol. Club, Cambridge, Mass.
Koplin, j. R. and R. S. Hoffman. 1968. Habitat overlap and comjrctitive exclusion in
\o\es ( Microt us ) . Am. Midi. Nat. 80:494—507
MacArthur, R. H. and J. W. MacArthlr. 1961. On Irird species diversity. Ecology
42:594-598.

Odum, E. P. 1969. The strategy of ecosystem development. Science 164:262-270.
Pearson, 1). L. 1971. Vertical stratification of birds in a tropical dry forest. (Condor
73:46-55.

SciioENER, T. W. 1974. Resource partitioning in (‘cological communities. Science 185:
27-39.

Shannon, C. E. 1948. A mathematical theory of communication. Bell Syst. Tech. J.
27:379-423, 623-656.

THE WILSON BULLETIN • Vol. 90, Nu. 1, March 1978

SiKGKi., S. I9S0. Nonparamctric statistics for the behavioral sciences. McGraw-Hill Co.,
New ^ ork.

Tkamkk, E. J. 1%9. Bird species diversity: components of !<hannon’s formula. Ecology
.SO: 92 7-929.

SCHOOL Of^ F0KP:STRY AM) WILDLIFE MANAGEMENT, LOUISIANA STATE UNIV.,
BATON KOUGE 7()8().'^ ( IMtESENT ADDRESS .JGD: SOUTHERN FOREST EXPERI-
MENT STATION, USDA FOREST SERVICE, NACOGDOCHES, TX 75961). AC-
CEPTED 31 MARCH 1976.

REQUESTS EOR ASSISTANCE

International shorehird surveys 1978. — A cooperative International Shorebird Survey
scheme has been organized by the Canadian Wildlife Service and the Manomet Bird
Observatory since 1974 to obtain information on shorehird migration and to identify and
document areas of major importance. This scheme has been highly successful, with much
very valuable information on shorehird distribution and migration coming from contribu-
tors throughout eastern Canada and the U.S.A., the Caribbean Islands, and Central and
South America. Information from the scheme will be valuable in assessing requirements
for the future protection and conservation of the birds and their habitat. It is planned
to make 1978 the fifth and final year of the project. Any observer who may be able
to participate in regular counts of shorehirds during spring and autumn migration periods,
as w'ell as during the winter in shorebird wintering areas, is asked to contact one of the
undersigned. Occasional counts from observers visiting shorebird areas on an irregular
basis would also be most welcome. For areas in Canada: Dr. R. I. G. Morrison, Canadian
\^’ildlife Service, 2721 Highway 31, Ottawa, Ontario. Canada KIG 3Z7. For areas in
U.S.A., Caribbean Islands. Central and South America: Brian A. Harrington, Manomet
Bird Observatory, Manomet, MA 02345.

Shorebird color-marking. — In 1978, the Canadian Wildlife Service will be continuing
a large-scale program of banding and color-marking shorebirds in James Bay. During
the past 3 years, over 30,000 shorehirds have been captured, resulting in more than 1200
‘‘bird days” of sightings of dyed birds ranging from eastern Canada to South America.
Much valuable information on migration routes and strategies is being obtained and ob-
servers are again asked to look out for and report any color-dyed or color-banded shore-
birds that they may see. Reports should include details of species (with age if possible),
place, date, color-marks and, if possible, notes on the numbers of other shorebirds present.
I’or color-dyed birds, please record the color and area of the bird that was dyed. For
color hands and standard metal leg hands, please record which leg the bands were on.
whether they were above or below the “knee,” the colors involved (yellow or light blue),
and the relative position of the hands if more than one was on a leg ( e.g. right lower
leg, blue over metal, etc.). All reports will be acknowledged and should be sent to:
Dr. R. I. (L Morrison, (Canadian Wildlife Service, 2721 Highway 31. Ottawa, Ontario,
(’anada KUi .3Z7.

AGRICULTURAL IMPACT OF A WINTER POPULATION OF
BLACKBIRDS AND STARLINGS

Richard A. Dolbeer, Paul P. Woronecki, Allen R. Stickley. Jr., and

Stephen B. White

The major concentration of blackbirds and Starlings (Sturnus vulgaris)
in North America occurs in the southeastern United States where an estimated
350 million Red-winged Blackbirds iAgelaius phoeniceus ) . Common Crackles
{Quiscalus quiscula). Brown-headed Cowbirds \Molothrus ater), and Star-
lings congregate in winter roosts ( Meanley 1971, 1975, 1977 ). An estimated
75-100 major (containing >1 million birds) roosts form in the southeastern
states each year.

Increasing attention is being given to many of these roosts because of
nuisance problems, reputed health hazards, and agricultural damage asso-
ciated with them. Although considerable effort has been directed toward de-
veloping methods for reducing roosting populations ( Lefebvre and Seubert
1970), field applications of such methods have met with considerable public
opposition ( Graham 1976 ) . L nfortunately, little effort has been directed to
ecological studies of the various roosting species during the winter months.

The objectives of this study were: (1) to document food habits, habitat

preferences and use, and general feeding and roosting behavior of the various
blackbird species and Starlings using a large winter roost; and (2) to under-
take a preliminary survey of the impact that this large roosting population
has on agriculture within a 40 km foraging radius of the roost.

STUDY AREA AND METHODS

The study area included a large concentration of blackbirds and Starlings that roosted
during the winter of 1975-76 about 7 km east of Milan, Tennessee, on the Milan Army
Ammunition Plant site near the Gibson-Carroll county border. The roost, bounded by
highway and pastures, was in a 21-year-old, 4.5 ha loblolly pine i Finns taeda) plantation
with little understoiy. The birds bad been roosting each winter at various sites on the
ammunition plant since at least 1969. The birds foraged primarily in Gibson County.
This roost has been a center of controversy since the winter of 1974-75 when the Army
attempted to reduce the bird population by spraying the roost with a wetting agent ( Rus-
sell 1975) .

Gibson County is a leading agricultural county in western Tennessee (Hobson 1976).
In 1975, over 48,000 ba were planted to soybeans; 14,000 ha to cotton; and 11,000 ba to
corn. Production of wheat, historically a minor crop in the county, increased from 3200
ha in 1973 to over 8000 ba in 1975. The county bad the largest bog and cattle population
in western Tennessee in 1973-75, averaging about 55,000 and 50.000 head, respectively.

Population numbers and species composition. — Between 1 November 1975 and 5 March
1976, 28 estimates of numbers and 18 estimates (at least once weekly) of the species
composition of blackbirds and Starlings roosting near Milan were obtained usually by 2

THE WII.SON BULLETIN • Vol. 90, No. I, March 1978

<)l)S(‘i \(‘is. \\ (* made population (*slimates 1)\ Block-counting (Mcanlcy 1965) birds in all
major fliglil lines as they left the roost. 'I'o estimate species composition, we stood under
the flight lines of departing birds in the morning or returning birds in the evening. At
b'ast 100 randotn binocular sightings were made for each estimate of s{)ecies composi-
tion; the first bird that cnter(*d the field of view was identified and recorded.

liird census fur habitat use. — Five automobile routes totaling 80 km were established on
secondary roads .5 to .'SO km from the roost, d be routes were in zones where major flight
lines from the roost have traditionally occurred and where comj)laints of damage to crops
have been most prevalent ( Bussell 1975).

On 24 days between .'SO October and .'S March, bird censuses were conducted on these
routes by 2 observers in 1 automobile starting 0.5-1 b after sunrise and ending before
15:30 ( U.'^T ) . The starting route and direction were randomly selected each day; the
remaining routes were run in the most exjjeditious order. Routes were driven at 15-45
km/b. The vehicle could be stoi)ped for up to 1 min if necessary to observe a flock of
birds through binoculars. Numbers, species, composition, and associated habitat types
were recorded for all observed groups of 2 or more Starlings and/or blackbirds.

The 12 habitat tyj)es were: (a) pasture fincludes cemeteries and unimproved pastures

with broomsedge) ; (b) cornfields and corn stubbie; (c) wheat; <d) legumes; )e) soy-
bean fields and soybean stubble; (f) feedlots (must include feeding apparatus or closely-
fenced livestock); (g) woodlots, forest, or brush; (b) buildings; (i) thoroughly plowed
(little evidence of previous crop type); (j) cotton; (k) fallow (weedy fields not in
cultivation or pasture previous growing season); and (1) miscellaneous.

We ran 1 to 4 habitat surveys monthly on the census routes from late October to early
March to determine the relative proportions of the 12 habitat types. We recorded the
habitat type on each side of the road at 0.16 km intervals. Thus, 1000 sample points
were recorded for each survey.

Food habits. — ^Between 14 November and 29 Februarv, usually once weekly, we col-
lected .50 to 75 birds by shotgun as they settled into the roost at dusk. Each bird was
identified, sexed, and weighed before stomach and esophageal contents were placed in a
vial containing 5% formalin.

For the analysis of food items, the formalin soluti«)n was drained from each vial and
the vial contents were i)laced in a drying oven (40°C) for several hours. Each sample
was then placed in a Petri dish containing five 1-mm dots symmetrically placed 2 cm
apart in the form of the center and end points of an “X.” The dish was shaken and
stopped at random and the food item resting on or nearest each dot was recorded. This
procedure was repeated 5 times per sample for a total of 25 recordings per sample. The
total for each food item (corn, wheat, sorghum, weed seeds, tree fruits, and insects) was
multiplied by 4 to obtain a ])ercentage estimate based on surface area.

This food-hahits analysis is biased toward foods that are difficult to digest fe.g., corn,
ceitain we(*d seeds) ; thus, the percentage estimates obtained should be viewed in light
of this limitation. In addition, birds were collected only in the evening; thus, any diurnal
pattern in food seb*cti\ity was ignored. Nonetheless, we believe this analysis provides a
general view of the dominant foods of the various bird species.

If aste corn. — Random areas of har\est«‘d cornfields in Gibson Ca)unty were searched
for corn at intervals during the winter to determine the amount of waste corn available
to birds. In each field, 2 strips 0.75 m by 55 m were searched. All cobs and pieces of
c(/bs bearing kermds were collect«‘d. In addition. 2 randomly-selected areas 0.75 m by
0.75 111 in each strip were searebed for loose kernels. I'liese loose kernels were eollected.

Dolbeer et al. • IMPACT OF BLACKBIRDS

Fig. 1. Estimated number of blaekbirds and Starlings using roost near Milan, Ten-
nessee, November 1975-early Mareh 1976.

and together with cob samples, removed from the fields. All kernels were removed and
weighed to the nearest gram. Kilograms of available corn per ha were then calculated.

Feedlots: bird censuses and observations. — On 9 days between 29 Januar>’ and 2 March
1976, we censused blackbird and Starling populations at 19 to 25 cattle and hog feed-
lots within 40 km of the Milan roost in the main area of foraging by the roosting pop-
ulation. Censuses were made from 10:00 to 15:00 CST. Each lot was observed for 5-10
min and the number and species composition of blackbirds and Starlings in the lot were
estimated. In addition, we made extended oliservations at several lots to study the feed-
ing behavior of the various birds species.

RESULTS

Size and species composition of roosting population. — The roost formed in
early November and disbanded in early March. The roosting population esti-
mates indicated a smooth growth to a peak population of around 11 million
blackbirds and Starlings in January and early February and a sudden re-
duction in late February (Fig. 1). The species composition remained fairly
constant between November and February (Table 1 j with an overall mean

THK WILSON lUILLKTIN • Vol. 90, No. 1, March 1978

'I'ahle 1

Avkkagk Monthly Si’KCiks Composition (%) of Birds at Milan, Tennessee Roost,
November 1975 Tiirolgii Early March 1976'

Early

Species

Nov.

Dec.

Jan.

Feb.

March

Mean

(iommon Crackle

70 (2.5)"

75 (5.2)

65 (6.7)

57 (6.3)

47 (0.2)

Red-winged

Blaekhird,

20 (0.7)

15 (1.0)

33 (3.5)

34 (.3.7)

4 (<0.1)

Brown-headed
Cowbird, and
Rusty Blackbird
.^tarling

10 (0.4)

10 (0.7)

4 (0.4)

9 (1.0)

49 (0.2)

^ Red-wiiiKS, cowbirds, and Rusty Blackbirds are lumped together because of our inability' to
distinguish during species composition estimates.

-Values in parenthesese are peak monthly population estimates (XlO’’).

of 64% grackles, 27% Red-wings and cowbirds, 9% Starlings, and a trace
% of Rusty Blackbirds i Euphagus carolinus). (Red-wings and cowbirds were
lumped together because they were difficult to distinguish during the species
composition estimates. Based on our general observations, the bird censuses
along roadsides, and the species composition of birds collected for food habits,
we believe that Red-wings were more numerous than cowbirds. I

Crackle populations peaked in January whereas Red-wing, cowbird, and
Starling populations peaked in February (Table 1). Crackle, Red-wing, and
cowbird populations decreased rapidly in late February before Starling pop-
ulations dispersed.

Daily habitat use. — The number of blackbirds and Starlings seen per census

Table 2

Number of Blackbirds and Starlings Recorded and Species Composition of These
Birds on Census Routes, Late Octobp:r 1975 Through Early March 1976, Milan.

Tennessee Area

No. days Species composition of birds observed (%) Avg. birds

census recorded per

Month

nin

Crackles

Red-w ings

Cowbirds

Starlings

Rusty BB

80 km census

Oct.

1565

Nov.

5430

Dec.

4537

Jan.

2404

Eel).

7622

March

2191

Total

UolbeeT et al. • IMPACT OF BLACKBIRDS

Table 3

Average % of Crackles, Red-wings, Cowbirds, and Starlings Observed by Habitat
Type for November 1975 Through February 1976, in Milan, Tennessee Area

Habitat

type

Avg. % of
total habitat
( Nov.-Feb. )

Crackles

Red-wings

Cowbirds

Starlings

Pasture

Corn

Wheat

Soybeans

Feedlots and legumes

Woods

Buildings and
miscellaneous

Plowed

Cotton

Fallow

Iluctuated considerably from day to day, primarily in response to weather
conditions ( e.g., during inclement weather more birds were usually seen ) .
The species composition of birds seen ( grackles 62%, Red-wings 17%, Starlings
16%, cowbirds 4%, Rusty Blackbirds <l%j was similar to the estimates ob-
tained at the roost ( Tables 1 and 2 ) .

The major habitat types along the census routes were pasture (23% ) and
soybeans (21%j followed by woods (13%), buildings and miscellaneous
( 12%), and fallow ( 11%) (Table 3). The relative proportions of the various
habitat types remained rather stable from November through February.

The various bird species had conspicuous differences in habitat use during
the winter months (Table 3). The majority of grackles were seen in corn-
fields and woods which made up only 4% and 13% of the habitat, respective-
ly. Red-wings were the only species commonly recorded in soybeans; 36%
of the Red-wings were seen in soybeans (21% of the habitat). Red-wings also
commonly used cornfields. We observed most cowbirds ( 65% ) at feedlots which
made up <1% of the habitat. Habitat types most used by Starlings were
pasture (31%), feedlots (23%), and woods (11%). They were also the only
species commonly seen around buildings and urban areas.

Some major changes in habitat use occurred during the winter (Fig. 2).
Starling use of wheatfields declined as the winter progressed. The use of
feedlots was rather constant throughout the winter for cowbirds; however, it
increased for Red-wings, Starlings, and grackles as winter progressed.

Food habits and body weights. — Corn (averaging 77%) was the dominant
food for grackles during each month, November through February (Table 4).

36 TllK WILSON lUJLI.ETIN • Vul. 90, No. 1, March 1978

•

l_ 10

_ /

Starlings in

Redwings in

■E

Wheat

Feedlots /

\ ^

'o 5

”* !

La-

1 1 ♦ ^ n

» — k'' 1 1

Nov. Dec. Jan. Feb.

• /

^ 50

(T3

’.5

Crackles

Starlings '

in •

Feedlots ,

Feedlots /

TZJ

/ •

! 25

■Si-.-rr" 1 0

L 1 i L

Nov. Dec. Jan. Feb. Nov. Dec. Jan. Feb.

Fig. 2. Percent of (A) Starlings observed in wheatfields, (B) Red-wings observed at
feedlots, (C) Crackles observed at feedlots, and (D) Starlings observed at feedlots during
bird censuses, November 1975 through February 1976, Gibson County, Tennessee.

Only in November, when acorns \Quercus spp. ) made up 29% of the food
items, was the percentage of corn <75% for grackles. The most common
weed seed found was ragweed {Ambrosia sp. I . Grackles, the only species
with sufficient numbers collected of both sexes to examine sex-specific food
habits, had no apparent sex-specific differences.

Corn ( 38% ) and weed seeds ( 36% ) were the dominant foods for Red-w ings
(Table 4l. Weed seeds commonly consumed were Johnson grass {Sorghum
halepense ) , cocklehur { Xanthium strumarium ) , chickgrass {Digitaria ischae-
mum), dropseed {S])orobolus sp. I , smartweed {Polygonum sp.), and pig-
weed { Amaranthus 9>\y.) .

Dolbeer et al. • IMPACT OF BLACKBIRDS

X? 1

1-0

'—

218

269

CNI

1— 1

•j'j

—

"S ■

CNI

ad 1

Urn

(Tj

:/]

c/5

^•

r-4

1— 1

C/5

;:i

1 £

' lO'

t_|

>■

r ^

C«

C-i

.-H

C/3

-0

c/3

Corn

Wheat

Sorghui

Tree se

Insects

inisct

Nunihei

Trace (<1%

THK WILSON lUILLKTIN • VoL <)0, No. I, March 1978

Iahle 5

Avkhagk Body Wkigiits in (^rams (±1 S.E.) for liLACKRiRDS Collkcted at Milan,
Tknnksske Boost, Novkmrkr 1975 Through February 1976

Montli

Male

Grackles

Female

Grackles

Male

Red-wings

November

126.1 ± 1.0

100.0 ±0.9

71.5 ± 0.6

(751*

(54)

(98)

Deeember

126.0 ± 1.7

101.1 ± 0.9

73.6 ± 0.5

(41)

(68)

Januaiy

132.3 ± 1.0

103.8 ± 1.5

79.0 ± 1.3

(17)

(14)

( 15)

Fel)ruary

126.4 ± 1.0

98.8 ± 1.2

72.8 ± 1.2

(64)

(26)

(24)

* Sample sizes are

in parentheses.

Starlings had the most cosmopolitan diet with plant seeds (29%), wheat
seeds or sprouts ( 19%), corn ( 11%), and tree fruits (9%) commonly present
(Table 4). Common plant seeds were sumac {Rhus sp.), wild grape iVitis
sp. ), and pokeweed [Phytolacca americana) . The tree fruits were predom-
inantly hackherry ( Celtis spp. ) . Starlings were the only species in which
insects were consistently found.

Corn (54%) and weed seeds (34%) were the dominant foods in the small
(15) sample of cowbirds collected (Table 4). Johnson grass, ragweed, and
pigweed were the dominant weed seeds.

Monthly average body weights ( November-February ) were determined for
male and female grackles and male Red-wings (Table 5). In all cases, weights
were highest in January. The average weight (±1 S.E.) for 13 female Red-
wings, 56 Starlings, and 9 male and 4 female cowbirds was 48.9 ± 0.4, 86.2 ±
0.9, 52.8 ± 0.6, and 40.0 ± 1.6 g, respectively.

IMPACT ON AGRICULTURE

Winter wheat. — A substantial part of our study was devoted to measuring
the damage inflicted by Starlings and blackbirds on winter wheat in Gibson
County. This study is the subject of a separate report ( Stickley et al. 1977)
and will he only briefly summarized here.

Forty-eight randomly selected wheatfields along the census routes were
monitored during the period November through February. Two of the 33
fields planted before 13 November received bird damage ( i.e., pulling up
sprouting wheat). All 15 fields planted after 12 November received some
damage. Seven of these 15 fields that could he assessed for damage averaged

Dolbeer et uL • IMPACT OF BLACKBIRDS

Table 6

Summary of Blackbird and Starling Numbers at Hog and Cattle Feedlots Censused
ON 9 Days From 29 January Through 2 March 1976 in Gibson County, Tennessee

Avg. no. Avg. species composition (%)

Type of Avg. no. lots birds in

lot censused/day lot/census Starlings Crackles Covvbirds Red-wings

Hog 14 54.0 (1000)“ 69 15 12 4

Cattle 8 223.0 (3000)“ 17 20 62 1

® Values in parentheses represent maximum number of birds estimated in a lot during a census.

about 11% of their sprouts removed by birds. Frost damage to the 15 late-
planted fields averaged 25% of the sprouts destroyed. Overall, the 48 fields
suffered an estimated 3.5% sprout removal by birds and 13.5% sprout destruc-
tion by frost. Almost no fresh damage was noted after mid-January. A sur-
vey of some of these fields in May 1976, shortly before harvest, revealed no
significant relationship between amount of bird damage in winter and num-
ber of mature heads of grain in May.

Almost all bird damage was done by Starlings (Table 4), a species that
made up a minor (9% ) part of the roosting population. Crackles commonly
were observed feeding in wheatfields (Table 3) ; but they fed on items such
as waste corn and weed seeds.

Corn. — Almost all corn was harvested in Gibson County by the time the
roost developed in early November. Thus, the roosting population inflicted
almost no damage to standing corn. However, both stomach contents (Table
4) and habitat-use censuses (Table 2) indicated that corn left in fields after
harvest was an important food for blackbirds, particularly grackles. Most of
this corn should be considered as waste corn, altough in some fenced fields
corn was used by livestock during the winter. About 17% of the cornfields
along the census routes were fenced.

The estimated waste corn per ha of harvested cornfields declined signifi-
cantly ( P < 0.05 ) from an average of 245 kg/ha ( moisture content of 20-
25% ) for 12 fields sampled in November to <10 kg/ha for 20 fields sampled
in January and February. Thus, far less corn was available to foraging black-
birds late in winter compared with November.

Feedlots. — There were major differences in blackbird and Starling popula-
tions in cattle and hog feedlots (Table 6). Cattle lots had an average of 223
birds per census compared to 54 birds for hog lots. Cowbirds, making up
62% of the birds observed, were by far the predominant species at cattle lots,
followed by grackles (20%) and Starlings (17%). In hog lots. Starlings,
constituting 69% of the birds seen, were by far the predominant species, fob

TIIK WILSON lUlLLKTIN • Vol. 90, No. 1, March 1978

lowed hy grackles (15%) and cowhirds (12%). Ked-wings made up <5%
of the birds at either cattle or hof>; lots. Although they were not enumerated
during feedlot censuses, House Sparrows (Passer dornesticus) were often as
numerous as Starlings or blackbirds. Blackbirds and/or Starlings were pres-
ent in about 75% of cattle and hog lots during the censuses.

d'he extended observations suggested that during midday, when the feed-
lot census was normally run, there was little exchange in the populations of
blackbirds and Starlings associated with the feedlots. In the early morning
and evening, however, there appeared to be a transfer of birds at feedlots
( i.e., new flocks coming in to feed as other flocks left ) .

d'he actual loss of feed to birds was not measured; however, some behavior-
al information on feeding and cjualitative information on feed loss was ob-
tained. In hog lots, almost no feed was lost to birds directly from feeders
because feeders were covered. Hogs commonly spilled grain out of feeders
and this was a source of food for the birds. Starlings, the predominant species
at hog lots, were the only birds seen perching on the hogs’ backs. The major
concern of hog farmers was the suspected role of the birds in the spread of
disease among hog lots.

In cattle lots, the predominant feeding activities of birds were: (1) feed-

ing on corn and other food items in cattle droppings, and (2) feeding in
pastures associated with cattle-feeding operations. Cowbirds, grackles. Star-
lings, and House Sparrows were observed in feeding troughs (covered feed-
ers are not used with cattle). Farmers expressed as much concern over feed
contamination from bird droppings as they did over feed consumption by
birds. Disease complaints were minor.

DISCUSSION

Niche Differentiation of the Bird Species

Grackles, Red-wings, cowbirds, and Starlings, although using a common
roost at night, had strikingly different niches in their daily existence.

Grackles. — Grackles were not only the most numerous species at the roost,
constituting about 64% of the birds and peaking at a population of over 6
million in January and early February; but, they also had the greatest body
weight per bird. Thus, if we assume a 1:1 sex ratio, grackles, from the view-
point of numbers, biomass, and energy consumed, had a greater impact than
the other species combined on the ecosystem w ithin the foraging range of the
roosting i)opulation.

lood-habits analyses and tbe habitat-use censuses showed that waste corn
was of primary importance to grackles. Overall, the negative impact that
grackles had on agriculture appeared rather minor. Areas of conflict were:

( 1 ) use of feedlots in late w inter, perhaps as a result of depleted supplies of

Dolbeer et al. • IMPACT OF BLACKBIRDS

corn in fields; and (2) competition with livestock feeding in harvested corn-
fields. Crackles were the most common species observed in wheatfields;
however, our food-habits data and observations indicated they fed on other
foods (often corn) and not wheat. Thus, their presence in wheatfields can
only be considered as beneficial. Thirty-one percent of the wheatfields had
previously been in corn and many had considerable amounts of corn at the
soil surface.

Red-wings. — Red-wings were the only species commonly associated with
harvested soybean fields (Table 3). A primary food source for Red-wings in
these fields appeared to be cocklebur, a common weed in soybeans. Corn was
also an important food source for Red-wings; they were often observed with
grackles feeding in corn stubble or wheatfields. Red-wings probably had the
least impact on agriculture of all roosting species. They, along with grackles,
increased their use of feedlots in late winter (Fig. 2), but were still a minor
species at feedlots ( Table 6) .

Cowbirds. — Cowbirds were the least common species in the roosting popula-
tion (with the exception of Rusty Blackbirds). They also had the lowest body
weight per bird and, thus, the lowest total biomass. They were primarily as-
sociated with cattle, either in pastures or feedlots. Corn, both from feedlots
and harvested fields, was an important food source.

Starlings. — Starlings made up 9% of the roosting population and about
8% of the biomass; they appeared to have the greatest negative impact on
agriculture of all species. They inflicted almost all the damage to sprouting
wheat and were by far the dominant species associated with hog lots. In
addition, they were the only species commonly associated with buildings and
urban areas. Starlings also used pastures and grassy areas much more than did
the other species.

Of considerable interest also is the fact that many of the Starlings roosting
in the mid-southern United States in winter (such as in Cibson Couny) are
year-round residents of the region. For example, B. L. Monroe ( pers. comm. )
has estimated that 47% of the Starlings roosting in Kentucky in the winter
remain in the state year-round. This contrasts with the vast majority of win-
ter-roosting grackles. Red-wings, and cowbirds that are migrants from the
northern United States and Canada (Fankhauser 1968, Meanley 1971, Mean-
ley and Dolbeer 1977, Dolbeer 1978) .

IMPACT OF BIRDS ON AGRICULTURE

Winter wheat. — The 48 wheatfields surveyed received an estimated 3.5%
sprout removal (range 0-34%) by Starlings during the winter and 13.5%
sprout destruction ( range 0-50% ) by frost. There was no apparent relation-
ship between amount of bird damage in winter and number of mature heads

THK WILSON lUiLI.ETIN • VoL 90, No. 1, March 1978

of prain in May. I luis, overall, bird damage to sprouting wheat appeared to
he a minor problem in the winter of 1975-76. Most bird damage and all
freeze-thavN damage occurred to fields planted after 12 November; therefore,
a preliminary management recommendation for areas in western Tennessee,
where bird damage is a problem, is that wheat should he planted before early
November when possible. ( Of course, additional data are needed to deter-
mine year-to-year and geographical variability before final management
recommendations are made. I For reasons other than bird damage, the Ten-
nessee Agricultural Extension Service recommends that wheat be planted be-
fore 1 November (Cobble 1974).

Feedlots. — This study documented that blackbirds and Starlings commonly
used feedlots, and that the feeding behavior and species composition of the
birds were different for hog and cattle lots. In both types of lots, the pre-
dominant species (Starlings or cowbirds) were species that made up a minor
part of the bird population at the Milan roost. House Sparrows, which do
not roost w ith blackbirds and Starlings and are present year round, were also
commonly found at the feedlots.

Additional (juantitative studies are needed to document: (1) the propor-

tion of feed consumed by birds that represents a real loss to the farmer ( Bes-
ser et al. 1968, Feare 1975), (2) the loss of feed in cattle lots due to con-
tamination by birds, and (3) the role that birds at feedlots play in the spread
of diseases. Also, more w ork, both extension and research, needs to be done
in the area of feedlot design and in the use of toxicants and mechanical scare
devices to reduce bird populations at feedlots (e.g., West 1968, Besser et al.
1967, Wright 1973 ).

Corn. — Standing corn received insignificant damage from the roosting
population because almost all was harvested by the time the roost formed in
early November. Nevertheless, waste corn was a dominant and perhaps critical
food item, particularly for grackles and Red-wings. Further studies should
be undertaken on the relationship of roosting populations to corn. The abun-
dance of waste corn in fields may be the major factor allowing many of these
large roosting populations to exist in their present locations throughout the
w inter.

CONCLUSIONS

ddie 1 bird species that composed the roost near Milan, Tennessee, varied
considerably in their numhers, biomass, foraging behavior, food habits, and
impact on agriculture. Because of this overall diversity and the complexity
of the various problems the birds create, simplistic management schemes are
likely to fail in solving the conflicts — they may even exacerbate them. For
example, much of the nuisance, feedlot problem, and wheat damage in Gibson

Dolbeer et al. • IMPACT OF BLACKBIRDS

County was caused by Starlings, a minor species in the roosting population.
Attempts to exterminate or greatly reduce the roosting population hy spray-
ing with a wetting agent may selectively favor the Starling, a highly adaptable
and prolific species that perhaps survives wetting-agent applications and low-
temperature stressing better than native blackbird species ( Odum and Pitelka
1939, Lustick and Joseph 1977) .

Long-term relief from the various conflicts most likely will require an in-
tegrated management program with a sound ecological basis. The use of bird-
control chemicals at feedlots, roost dispersal and roosting habitat manipula-
tion techniques, lethal control at certain roosts, changes in certain cultural
practices in agriculture, chemical repellents, and public tolerance may all be
a part of such an integrated approach. We hope that this preliminary study
has contributed information that will help make such a management program
a reality.

SUMMARY

Hal)itat preferences and use, food habits, and impact on afirieulture were studietl for
11 million l)lackbirds and Starlings roosting in Gibson County, Tennessee, in the winter
of 1975-76. The roost was composed of Common Crackles (64%), Red-winged Black-
birds and Brown-beaded Cowljirds (27%), Starlings (9%), and Rusty Blackbirds (<!%)•
The various species bad strikingly different niches in their daily existence and impact on
agriculture. Waste corn was of primary importance to graekles and Red-wings, which
spent most of their time feeding in corn and soybean stubble and woodlots. Red-wings
fed commonly on cocklebur seeds in soybean stubble. Cowbirds and Starlings commonly
used pastures and feedlots. Starlings did almost all the bird damage to sprouting wheat.
Starlings, a minor species in the roost, bad the greatest negative impact on agriculture.
Simplistic management schemes are likely to fail in solving bird-man conflicts caused by
multi-species roosting populations — they may even exacerbate them. Long-term relief
most likely will re(juire an integrated management program with a sound ecological
basis.

ACKNOWLEDGMENTS

We thank P. W. Lefebvre, J. C. Rogers, Jr.. B. Meanley, W. C. Royall, Jr., and I).
A. Buecker for assistance in the fieldwork. T. McCutcben, Milan Field Station. Univer-
sity of Tennessee, generously cooperated with us in many ways. We thank personnel of
the Milan Army Ammunition Plant, particularly Colonel Percbutz, W. Oates, and S.
Stevenson, for their interest and coo])eration.

LITERATURE CITED

Besser, J. F., W. C. Royall, Jr., and J. W. I)e Crazio. 1967, Baiting Starlings with
DRC-1339 at a cattle feedlot. J. Wildl. Manage. 31:48-51.

— , J. W, De (iRAZio, AND J. L. CuARiNo. 1968. Costs of wintering Starlings and

Red-winged Blackbirds at feedlots. J. Wildl. Manage. 32:179-180.

Cobble, R. E. 1974. Growing wheat in Tennessse. Agric. Exten. Serv., Univ. Tenn.
Publ. 576 (Rev).

THE WILSON BULLETIN • VoL 90, No. 1, March 1978

Doujkkr, IL a. 1978. Movement and migration patterns of Ked-winged Blackbirds: A
contimmtal overview. Bird-Banding I In press. 1
Ea.nkii Ai SKK, 1). I’. 1968. A comparison of migration between blackbirds and .Starlings.

Wilson Bull. 80:22.V227.

Ekakk. C. .1. 1975. (iost of Starling damage at an intensive animal husbandry unit.

I’roc. Br. Insecticide and Fungicide Conf. 8:2.53-2.59.

Uhaham, E., .1h. 1976. Blackbirds — A problem that won’t fly away. Audubon 78:118-

125.

Hobson, K. 1976. Tennessee Agricultural Statistics. Annu. Bull. 1976. Tenn. Crop
Kej). Serv. Bull. T-13.

Lkfkbvre, \\ W., AM) .1. L. Sei;bp:rt. 1970. Surfactants as blackbird stressing agents.

Proc. Vertebr. Pest Conf., Univ. Calif., Davis 4:156-161.

Lustick, S., and a. Joseph. 1977. Seasonal variation in the effects of wetting on the
energetics and survival of Starlings (Sturnus vulgaris). J. Comp. Biochem. & Physiol.
56:173-177.

Meanley. B. 1965. The roosting l)ehavior of the Red-winged Blackbird in the southern
United States. Wilson Bull. 77:217-228.

. 1971. Blackbirds and the southern rice crop. U.S. Dept. Int., Bur. Sport Fish.

Wildl. Res. Publ. 100.

. 1975. The blackbird-Starling roost problem. Atl. Nat. 39:107-110.

— . 1977. The 1974-75 winter roost survey for blackbirds and Starlings. Proc.

Bird Control Seminar, Bowling Green State Univ., Bowling Green, Ohio 7:39-40.
, AND R. A. Dolbeer. 1977. Source of Common Crackles and Red-winged Black-
birds wintering in Tennessee. Migrant 48: fin press.]

Odum, E. P., and F. A. Pitelka. 1939. Storm mortality in a winter Starling roost.
Auk 56:451-455.

Russell, H. G. 1975. Blackbird control on two Army installations — Environmental
Impact Statement. Office Chief, Eng., Directorate Facilities Eng., Wash., D.C.
f A copy has been placed in the Van Tyne Library, Univ. of Mich., Ann Arbor.)
Stickley, a. R., Jr., R. A. Dolbeer, and S. B. White. 1977. Starling damage to
sprouting winter wheat in Tennessee. Proc. Bird Control Seminar, Bowling Green
State Univ., Bowling Green, Ohio 7:30-38.

W EST, R. R. 1968. Reduction of a winter Starling population hy baiting its preroosting
areas. J. Wildl. Manage. 32:637-640.

Wright, E. N. 1973. Experiments to control Starling damage at intensive animal hus-
bandry units. OEPP EPPO Bull. 9:85-89.

U.S. FISH AM) WILDLIFE SERVICE, OHIO FIELD STATION, DENVER WILDLIFE RE-
SEARCH CENTER, C/O PLUM BROOK, TAYLOR AND COLUMBUS ROADS, SAN-
DUSKY, OHIO 44870.

(present address, SBW: cooperative wildlife research unit, OHIO STATE
UNIVERSITY, COLUMBUS, OHIO 43210 ) . ACCEPTED 31 JULY 1977.

BREEDING BEHAVIOR OE THE LOUISIANA HERON

James A. Rodgers, Jr.

Information on the breeding behavior of the Louisiana Heron ( Hydranassa
tricolor ) lacks sufficient detail to permit comparison with other ardeids. Both
Bent (1926) and Palmer (1962) summarized the available data from the
literature and drew upon previously unpublished information from contribu-
tors. Other studies (i.e., Teal 1965, Jenni 1969) deal with the reproductive
success or ecology. I have described the displays characteristic of the breed-
ing period ( Rodgers 1977 ) . This paper reports on the breeding behavior and
nesting ecology of the Louisiana Heron.

STUDY AREA

Most fieldwork was carried out on Grand Island, Barataria Bay, Plaquemines Parish,
Louisiana. Except for a few scattered oyster shell beaches, the island is covered by 5-20
cm of water during high tide, hut drains almost completely at low tide except for a few
inland pools. Predominant vegetation includes black mangrove (Avicennia nitida) and
cordgrass (Spartina alterniflora) .

Observations were made from blinds at 2 sites. Site A was characterized by tall, dense
black mangrove averaging 1.5-2 m high with little or no open cordgrass areas. Site B
was characterized by numerous small, low-lying clumps scattered throughout an equal
area of cordgrass. Herons were designated as to which site they occupied, for example,
male A1 or Bl. Most herons and ibises fed in surrounding estuarine regions of Barataria
Bay. Hence, Grand Island was used by the Louisiana Heron primarily as a breeding
site, though also used for roosting at other times of the year.

PLUMAGE CHANGES

Palmer (1962) described the nuptial molt in preparation for breeding by
the Louisiana Heron. The soft-part color depicted in a colorplate (page 367)
that accompanies the description retiuires some revision for the Louisiana
Heron on the coast of Louisiana.

During the courtship phase of the breeding season, the bill tip was black
as indicated, hut the remaining % of the bill, including the orbital skin, was
turquoise-cobalt (terminology of Ridgway 1912). Displaying males fre-
quently rubbed the bill over the oil gland during preening with the result
that the soft part colors became more intense. As noted by Huxley (1922),
the turquoise-cobalt color was typically less intense in females. I he entire iris
of males was more of a magenta color; in females the inner margin of the iris
around the pupil turned to rose. Leg color of both sexes was fleshy maroon.

During the incubation period, the turquoise-cobalt color changed to a
mottled yellow in both sexes. The color change occurred first around the

1^)

THE WILSON lUlLLETIN • VoL 90, No. 1, March 1978

eye. then alonji; the resl of the hill. The sexes could then no longer he dis-
tinguished. riie magenta color of the iris faded, the inner margin around the
pupil hecoming almost pink in both males and females. The legs l)ecame
grayish-yellow.

TEItKITOllY

Louisiana Herons exhibited no pairing behavior outside the breeding sea-
son and behaved as solitary individuals while foraging and roosting in the
colony. In February a slow 2-fold increase began in the roost population on
Grand Island until the heronry swelled to l)reeding density. Louisiana Herons
started to show soft-part breeding color and stayed longer on the island in
the morning or returned earlier in the afternoon. Males began spending much
time in a certain region of the island that eventually served as their territory.
Occasionally a male picked up and manipulated a twig and then dropped it,
or tugged on a branch. TTiese actions marked the onset of the breeding season
as the males began to set up their territories.

The male set up the territory, often built the foundation of the nest, and
undertook the initial defense of the territory from both intruding males and
females, as is the case in most North American ardeids ( Meanley 1955, Mey-
erriecks I960, Palmer 1962 ). Early male Louisiana Herons staked out large
territories with centers far apart; hut, as other males began to enter the breed-
ing cycle, the territories gradually became closer and territorial disputes over
boundaries became increasingly common. Once a male had selected a site, he
preened for long periods. Any disturbance in the vicinity or another Louisiana
Heron flying by was viewed intently. Soon after the male had chosen a ter-
ritory, he began to grasp and tug at branches around the site. A few low-
intensity performances of the Snap-Stretch display I Rodgers 1977 ) sometimes
occurred between long preening sessions. The male occasionally went below
the nest bush and returned with a twig that he used in the foundation of the
nest. No matter how long the male displayed, the nest never got past a loosely
arranged foundation stage.

Nests generally did not last until the next season so even the earliest males
had to construct new foundations when setting up a terriory. However, they
often used the stunted and leafless area amidst the branches where a nest existed
the previous season. Abandoned nests were almost always used again by a
male later in the season. Nest height ranged between 0.3-1. 3 m above the
high water mark. "1 he early males tended to set up a territory in the dense
growth of 1.5-2 m tall mangroves, while the later males nested in the lower,
more scattered mangroves.

1 he initial territory set up by the male Louisiana Heron ranged in size from
4 to 10 m in diameter. Between preening sessions and ever increasing Snap-

Rodgers • LOUISIANA HERON BEHAVIOR

Stretch performances, the male roamed his territory, occasionally stopping
to preen or display at a site other than the core display site. Initially all herons
were threatened upon their approach to the territory and attacked if they did
not leave. Pursuit flights were very frequent during this period of the breed-
ing cycle. In general, an intruder was intimidated by the threat display of a
resident male and proceeded no further. Once boundaries were established,
the territory was recognized and seldom contested; hence aggression between
neighboring males decreased with time.

As females were gradually allowed to stay on the territory and an increasing
number of new males entered the heronry and began setting up their own ter-
ritories, the size of the initially large territory gradually shrank. The terri-
tory reached its minimum size just after pair formation. This phenomenon
is also characteristic of many other ardeids ( Meanley 1955, Cottrille and
Cottrille 1958, Meyerriecks 1960, 1962 ). If the nest of the Louisiana Heron
was located in one of the low, isolated bushes, the territory sometimes shrank
to include only the bush and its immediate vicinity, a diameter of 2-3 m. If
the nest was situated in large, dense mangrove, the territory included all the
hushtop area above the nest.

Figure 1 shows the history of one region of site A beginning on 11 March
1973, with special reference to male Al. Male A1 displayed from numerous
sites, but concentrated display activities on the core mangrove hush. In the
late afternoon of 14 March, male Al paired. From this time on, both male
and female participated in the defense of the territory. By 15 June, all nests
contained either eggs or young nestlings, with the exception of the aban-
doned A6 site. Male A6 allowed a female to come onto his nest, where 1
copulation took place. Afterwards, he drove her from the nest and did not
allow her back on. Male A6 never paired.

In the large heronry on Grand Island, territory establishment seemed to be
staggered. Because the territory of a given pair shrank and left undefended
areas, other males moved in and sometimes began displaying as close as 2 m
from an incubating heron. These new males were noticeably tense during this
period, exhibiting sleeking and constant head turning as they observed the
surrounding heronry. With time, they began to roam the mangrove hush,
displaying from any site where they would not be threatened or attacked by
neighboring herons. Soon a core display site developed, which was usually
the place where the nest foundation was begun.

As the courtship period progressed, male Louisiana Herons performed their
Snap-Stretch and Circle Flights (Rodgers 1977) from the nest, hut often
could be observed moving to another site and displaying from there. If the
nest or core site was located deep down in the mangrove bush, the male oc-
casionally moved to the hushtop or out on a limb to display, making him

THE WILSON nULLETIN • Vol. 90, No. 1, March 1978

Fig. 1. The history of territory formation in one region of site A during 1973. During
the early morning of 11 March, male A1 was alone and displayed from numerous sites
designated hy open circles, but concentrated display activities on the core mangrove bush
designated with a solid circle. Later in the day A2 appeared. Territorial boundaries are
indicated hy dotted lines and were determined hy mapping display sites and the locations
of aggressive clashes.

more conspicuous to Loth roaming females and other territorial males. Ter-
ritorial male Louisiana Herons oriented the head and looked in the direction
of the acoustic cues produced in the Snap-Stretch. After pair formation,
the female joined in the defense of the territory, a pattern typical of many
other ardeids ( Meyerriecks 1960, 1962, Blaker 1969j. Later when the nest-
lings were 3 weeks and older they too defended the nest from both adult and
juvenile ardeids when their parents were away.

Certain points should he brought out in regard to the shrinking effect of
the male’s territory. First, a large territory at the early stage of courtship al-
lowed more display sites and functioned to make the male more conspicuous
to females wandering through the heronry. Second, a larger territory had
more area from which the female could observe the displaying male without
being driven off immediately (Meyerriecks I960). Third, when the male fi-
nally accejited the female and formed a pair bond, a large territory was no
longer needed. As a result, more males could come in and occupy newly un-
contested areas of the heronry and the high breeding density characteristic of

Rodgers • LOUISIANA HERON BEHAVIOR

MALE TWIG
SHAKES

MALE THREATENS
FEMALE

MALE ATTACKS
FEMALE

MALE AND FEMALE
BILL-NIBBLE

FEMALE-NO PAIR
BOND FORMED

FEMALE LANDS
ON TERRITORY

FEMALE ALLOWED
TO STAY

FEMALE MOVES
CLOSER TO CORE

FEMALE ASSUMES

MALE PERFORMS
SNAP AND STRETCH

MALE PERFORMS
" CIRCLE FLIGHT

FEMALE PERFORMS

WITHDRAWN CROUCH GREETING DISPLAY

t /

FEMALE ALLOWED
CLOSER OR
ONTO CORE

MALE AND FEMALE
PREEN OR

STABLE PAIR
BOND FORMED-
NESTING BEGINS

Fig. 2. Possible intercations between male and female Louisiana Herons during the
courtship period. P"or a full discussion see the text.

this species developed. Fourth, the decrease in the territory size permitted the
pair to channel their energy away from territory defense and into reproduc-
tion. Fifth, because the initial nest foundation served as the core for display-
ing and later as a site for raising the young, the immediate area around the
nest was defended most vigorously at all times.

INTERACTIONS WITHIN THE PAIR

Male selection. — There are many possible interactions between males and
females during the courtship period (Fig. 2). Females at first landed on the

I HE WILSON lUJLI.KTlN • Vul. 90, No. 1, March 1978

edf>;e of the displayiiif? male’s territory. Showing little overt interest in the
displaying male or the nest, the females at this time usually did not stop for
any great length of time and seemed to he just moving through the colony.
If the female moved too far into the territory of the male, he might first ex-
hibit Luig Shaking (Rodgers 1977) or immediately drive her away as he
would an intruding male. During the early part of the breeding season fe-
males generally did not return when driven away by the male. Later the per-
sistent female would return repeatedly despite repeated threats and attacks by
the male. Male Louisiana Herons during the early phase of courtship per-
formed occasional Circle Flights and Snap-Stretches in between long periods
of preening or twig manipulation (Fig. 3). The preening exhibited by males
during this period was often marked by rapid, forceful, erratic movements.
Later, preening movements were more relaxed. With an increase in time, the
male performed a greater number of Snap-Stretches I Fig. 3).

At the approach of a female Louisiana Heron, the male generally returned
to the core site to perform Snap-Stretch and Circle Flight displays. Once a
female became attentive to a displaying male or his nest, she got as close as
possible. At first she was thwarted in her attempts to get closer than the edge
of the male’s territory. Even though performing Snap-Stretches, the male at
this time showed predominantly aggressive behavior. Twig Shaking was com-
monly observed as he threatened the female and often drove her off in pur-
suit flights. A persistent female was allowed to remain on the edge of the
territory, only eliciting mild feather erection in the male. Continual attempts
by the female eventually resulted in her getting within a short distance of the
nest. During this time, the male repeatedly did Twig Shakes. He threatened
the female with Lpright and Forward displays (Rodgers 1977) and oc-
casionally drove her back some distance or even off the territory. Often the
female remained in view after being chased off by the male and returned onto
his territory repeatedly, especially during the late stages of the courtship phase.
Meyerriecks (1960) described similar behavior on the part of the female
Green Heron { Butorides virescens ) .

The presence of a female initially increased the fre(iuency of Snap-Stretch
and Circle Flight displays, which were occasionally performed in rapid suc-
cession ( Fig. 3 ) . The maximum number of displays observed during any 1
min ( N = 8328 min ) w as 5 Snap-Stretcbes and 3 isolated Snaps performed
by male A1 with a female on the territory.

I he female intently observed the male displaying, preening, and arranging
twigs in the nest. Her attention seemed to be directed not only to the male,
hut the nest foundation if present. W hen the male performed the Circle Flight,
the female watched the entire performance. Females were often observed
jumping onto the nest while the male was engaged in the Circle Flight. Upon

Rodgers • LOUISIANA HERON BEHAVIOR

I II III IV

1687 min 2672min 2449 min 1520 min

Fig. 3. Shifts in the hehavioral repertoire of male Louisiana Herons, Phases of the
courtship period are: L first L2 h of displaying; II, after first h of displaying; HI,
behavior with a female present; IV, last Y2 h before copulation. Abbreviations are: A,
aggressive behavior (includes the Upright, Forward, and Twig Shake displays); BN,
Bill-nibble; CF, Circle Flight; GI). Greeting display; P, preening; S, Snap; SS, Snap-
Stretch; TM, twig manipulation.

his return, the male always attacked the female and drove her away. Oc-
casionally, as many as 3 females joined the male in the Circle Flight by flying
after him and landing nearby when he returned to his nest. The female’s
flight did not in any way resemble the initial components of the Circle Flight
of the male, but the return flight was similar in posture and calls to the Greet-
ing display phase performed liy the male ( see Rodgers 1977 j .

THK WILSON IUjLLETIN • Vol. 90, Nu. I, March 1978

1 he female was allowed closer and closer until she was within 1-1.5 m of
the nest. She attempted to get even closer and assumed the Withdrawn Crouch
(Rodgers 1977). During an encounter with a threatening or attacking male,
female Louisiana Herons often i erformed elements of the Greeting display,
which aj)peared to function in reducing the aggressiveness of the male (Rod-
gers 1977). Often the male stopped his attack and temporarily joined in the
performance of the Greeting display. This outcome allowed the female to hold
her position or move slightly closer to the nest.

As the female assumed the Withdrawn Crouch, she frequently Bill-nibbled
(Rodgers 1977). If she rose from her low-profile posture, she was attacked
and driven off. hen the female w as on or near the edge of the nest, the male
exhibited a tendency toward a reduction of Snap-Stretch and Circle Flight
displaying (Fig. 3). In addition, much preening and Bill-nibbling was done
by the female, with a slow increase in the frequency of Bill-nibbling by the
male ( Fig. 3) .

Precopulatory period. — This period of courtship from the time the female
is first tolerated on the nest to just before the first copulation by the po-
tential pair was marked by a sw itch from aggressive to sexual behavior in the
male. For long periods both male and female engaged in preening, interrupted
by twig manipulation and Bill-nibbling. This appears to be typical of pre-
copulatory behavior of most ardeids ( Meanley 1955, Cottrille and Cottrille
1958, Meyerriecks 1960, 1962, Blaker 1969 ) . The male Louisiana Heron still
exhibited moderate feather erection, while at the same time Bill-nibbling be-
came more frequent. He rarely performed the Snap-Stretch and Circle Flight
displays once the female was on the nest I Fig. 3 ). The 2 herons often stood
side by side, usually facing in opposite directions for periods up to 40 min.
The female then rose out of the Withdrawn Crouch and stood with the male
in the center of the nest w ithout being attacked.

Copulation. — Soon after the female succeeded in getting onto the nest, the
pair copulated either on (36 of 39 observations) or close to the nest platform
or core display site. The male watched the female as she began manipulating
tw igs in the nest, and they both often Bill-nibbled. The female leaned forw ard,
withdrew the head partially, and s(iuatted. This act probably indicated to the
male that she was ready to copulate. I observed no precopulatory display.

1 he male mounted the female from the rear or the side by stepping up onto
her back. Flapping his wings to maintain his balance, he sciuatted on her,
grasping her shoulders with his toes. He also used leverage of his bill on the
shoulder and neck region of the female for additional support. Meanley
( 1955) reported similar use of the bill by the male Little Blue Heron {Florida
cacrulea). Copulations ranged in duration from 8 to 11 sec (mean 9.8 sec;
N = 23). No feather erection, nor any calls were noticed at any time pieced-

Rodgers • LOUISIANA HERON BEHAVIOR

ing, during, or immediately after copulation. Finishing copulation, the male
rose and stepped off the back of the female. Copulation seemed to occur ir-
regularly during the early phase of pair formation, though I observed copula-
tions throughout the egg-laying period.

The length of time a male maintained a territory and engaged in courtship
behavior without pairing is not accurately known. One instance recorded was
the occupation of a territory by male A6 for only 4 days before abandoning
it. Male All maintained a territory for 11 days before pairing.

Although I did not mark adult Louisiana Herons, I believe they practice
seasonal monogamy. In 2 years, I observed no evidence of promiscuity once
the male and female had paired, nor ‘“rapes” as have been reported for the
Little Blue Heron ( Meanley 1955 I .

Postcopulatory period. — Immediately after the first copulation, the pair
members engaged in long periods of preening. Bill-nibbling, and manipulating
nest twigs. They also spent a lot of time standing side by side, facing in opposite
directions. In this position, Huxley (in Bent 1926) has said they “intertwine
their necks,” but this is actually an illusion produced by their relative posi-
tions. The female soon began rearranging the twigs in the nest with the male
looking on. Occasionally, both herons suddenly engaged in short perfor-
mances of the Greeting display.

If the female left the nest after the first copulation to preen or find twigs,
the male might not allow her back onto the nest. I observed 4 instances ( N =
39 copulations ) in which a male copulated with a female and then attacked
her and drove her off or did not allow her return onto the nest after she had
left it momentarily. This led to dissolution of the bond in 2 instances when
the male repeatedly repulsed the female’s attempts to get back on the nest.

Males under observation did not feed while courting. The presence of wan-
dering females and intruding males may be the reason for their abstinence.
Only after pair formation, with the presence of its mate to defend the terri-
tory, could the male forage without danger of losing his territory. Jenni
( 1969 ) also reported that male Louisiana Herons do not feed during the court-
ship period.

^EST BUILDING

I he male was usually the one who went out and found twigs. Male Louisi-
ana Herons passed these twigs to the female, who then worked them into the
nest. Typically the period from the first copulation to the initiation of stick
collecting was short. For example, pair A9 copulated at 11:43 and the male
returned with the first twig in 11 min. Most twigs were gathered from on
or very near the territory, hut occasionally males brought material from
greater distances. Birds would also take twigs from abandoned nests nearby.

When the male returned with a twig in his hill, the pair performed the

rHK WILSON miLLPyriN • VoL <J0, Nu. I, March 1978

rri

(ireelinfi display ( l{o(lj>:ers 10771. I{eachiiif>: out, the female took the twip; in
her hill and hotli engaf>:ed in additional head noddirifj; and calling. The fe-
male then placed the twig in the nest as the male looked on. I he male often
Bill-nihhied while the female v\as inserting the twig. Initially, the male ex-
hibited considerable feather erection while passing twigs to his new mate.
This feather erection decreased as nest building continued and was limited
to the crest and, to a lesser extent, the aigrettes.

I he nest foundation begun by the male during the courtship period con-
sisted of large tw igs about 1 cm in rliameter and 30-60 cm long ( N = 27 I .
Phis structure sometimes had only 2 or 3 twigs. Twigs brought to the female
after pair formation, while occasionally large, usually were about 0.5 cm in
diameter and shorter than 30 cm (N = 163). The finishing material con-
sisted of small, flexible twigs and Sj)artina grass stems. The completed nest
varied in w idth and depth, hut always had a slight depression on the top sur-
face that j)re\ented the eggs from rolling out. I could distinguish no differ-
ences between the nests of Louisiana Herons and those of Little Blue Herons
or Snow y Egrets ( Egretta thula I on Grand Island.

During nest building, one member of the pair was always on or near the
territory. This prevented neighboring herons from stealing nest material.
After the nest was completed, both sexes occasionally added twigs to the
structure for repairs between periods of incubating or brooding. The inter-
val between the first copulation and the laying of the first egg for 2 pairs of
Louisiana Herons on Grand Island w as 4 and 6 days. Jenni ( 1969 ) recorded
l)eriods of 4 and 5 days at Lake Alice, Florida.

INCUBATION STAGE

During the interval before the eggs were laid, the female spent a great
amount of time scjuatting down on the nest, rearranging the nest twigs, and
placing in new material. On Grand Island, the clutch size averaged 3.0 eggs
( range = 2-5, mode = 3 j . Of the 99 nests examined, 22 contained tw o eggs,
58 three eggs, 18 four eggs, and 1 had five eggs. For Lake Alice, Jenni ( 1969 I
recorded the average clutch size for 35 nests as 4.1 eggs (range = 3-9, mode
= 4l. deal ( 1965 I reported an average of 3.1 eggs per clutch for 15 nests
( range = 2-4, mode = 3 ) for Sapelo Island, Georgia.

After the clutch was completed, the pair was usually together only during a
nest relief. Both sexes incubated. Nest reliefs were irregular and were ac-
companied by tbe (/leeting disi)lay and Bill-nibbling. I he returned heron
usually preened or manipulated a few twigs in the nest before settling. Its
mate often i)reened a short time on the edge of the nest before leaving to col-
lect 1 or more twigs which were i)assed in a Greeting display to its mate who
inserted them into the nest. Huxley Un Bent 1926) stated the number may be

Rodgers • LOUISIANA HERON BEHAVIOR

as many as 11 twig presentations at 1 nest relief. My data indicate 2-5 pre-
sentations (mean 4.2, N = 167) are more common. Because the nests are in
good condition structurally, I believe the display functions to reinforce the
pair bond in these instances. Sometimes an incubating heron did not imme-
diately rise off the nest when its mate returned, but remained sitting on the
eggs. The returned heron then engagd in additional Greeting displaying and
Bill-nibbling, after which the mate would rise off the nest.

The length of an incubation period varied considerably, from periods of
less than 1 h to 1 recorded instance in which the partner did not come hack
during the daylight hours ( about 06:00 to 20:30 I . Such long periods between
changeovers were rare, and although nest reliefs were irregular, there tended
to be 1 during the mid-morning, another during the mid-afternoon, and often
1 during the early evening. Incubation was most continuous in the early
morning and early evening periods, when the ambient temperatures were cool-
er. During these periods the adult only occasionally rose off the eggs to
preen, rearrange nest material, or sunbathe. During the late morning and
afternoon, when the temperatures were the warmest, the Louisiana Heron often
rose off the eggs. The bird rarely left the eggs unattended; usually it stood
on the nest edge while preening or sunbathing. I never saw an incubating bird
leave the nest to forage in nearby shallow inland pools. As the heron sat again
on the eggs, it first erected the feathers of the lower breast and belly, then
dropped the wings slightly, and shifted from side to side as it lowered its
body onto the eggs. Mild crest erection also occurred.

Incubating Louisiana Herons were quick to rise and threaten any species
of heron intruding on the territory. The area immediately around the nest
was defended most vigorously. Disturbances outside the territory usually
elicited only the Alert Posture ( Rodgers 1977 ) .

At pipping, the young Louisiana Heron emits high-pitched “pcepmg”
sounds. During this time, adults looked at the eggs for long periods of time,
sat, then rose again and peered at the eggs while turning the head from side to
side. The adult heron later dropped pieces of eggshell out of the nest.

BROODING AND CARE OF THE YOUNG

Because incubation begins with the first egg, hatching is asynchronous.
Lntil the nestlings were about 1 week old, the behavior of the adults was the
same as during the incubation period, except for feeding the young. Adults
brooded most intensively during the cool periods of morning and early eve-
ning. Herons commonly sunbathed while sciuatting on the nest or standing
over and shading the nestlings.

From age 1 day to about 1 week, nestlings were fed small fish that were
regurgitated by the parent onto the floor of the nest. The nestlings picked

TIIK W ILSON lUiLLK I’IN • Vol. 90, No. 1, March 1978

up the fish in their hills. Similar hehavior has been reported for the Little
Blue Heron ( Meanley 1655), Great Blue Heron, Arden herodias ( Pratt 1970 ),
and (5ittle Lfjret Hulbulcus ibis (Weher 1975). Afterwards the adult ate the
unconsuined fish. I pon nest relief, the departing parent often fed the nest-
lings 1 more time.

W hen the young were 1-2 weeks of age, the hill, legs, and general body
strength were well enough developed to allow grasping of the parent’s hill and
they obtained a meal directly from the adult. The young were then fed one
at a time with the adult leaning forward and extending the mandibles down-
ward in such a manner that the dorsal surface of the upper mandible faced
the offspring and was grasped. After feeding one nestling, the adult usually
moved a short distance from the calling chicks before returning and feeding a
second. Ibis i)iocess was usually repeated several times. In such sessions, 1
nestling was often fed twice. Adults stood on the edge of the nest and could
easily elude clamoring 1-2-week-old chicks; older young chased the parent
around the nest hush. Hence, between feeding bouts the adult flew a short
distance away to rest, preen, or sunbathe.

After the nestlings were 3-4 weeks old, the parents were seldom seen with
the young except when feeding them. Pratt (1970) reported parent Great
Blue Herons stay away after the young are 28 days old; W^eber (^1975) found
that 14-21 -day-old Cattle Egret chicks are left by themselves. Perhaps as the
young Louisiana Herons got older, almost constant effort on the part of both
})arents was recjuired to supply the developing juveniles with an adequate
(juantity of fish. Siegfried ( 1972 ) found that Cattle Egret nestling growth
curves are steenest (hence high food demands), and chick mortality due to
starving is highest at the time when both parents switched from alternate to
simultaneous hunting regimes. Also, young Louisiana Herons at this age
were old enough to defend themselves. With 4-5-week-old juveniles, the par-
ent did not land on its nest hush hut landed 2-5 m away, gave a series of
calls and waited for its offspring to come to it. During this time, the
parent exhibited moderate feather erection which increased slightly as the
young approached to receive food. As food was transferred directly to the
juvenile, both the parent and young herons rapidly flapped their wings to
maintain an upright position. After feeding its brood, the parent stepped and
faced away, then reswallowed any fish that remained in its mouth.

Because parent Louisiana Herons refused to feed many begging juveniles
that approached them, I suspect that they were able to recognize their young,
(/enerally, the young 3-4 weeks old were the ones that approached nonparents
for food. I he older juveniles apparently learned to recognize their parents
and probably by the process of habituation ( in the terminology of Thorpe
1963), they learned that they would not get food from hut would he at-

Rodgers • LOUISIANA HERON BEHAVIOR

tacked by a nonparent heron. Skead ( 1966 ) noticed adult-young recognition
in the Cattle Egret.

The parents returned at irregular times to feed their young. The number
of feedings was usually 4 or 5 per day. When the juveniles were 7-8 weeks
old, the parent-young bond had disintegrated and the juvenile herons were
on their own. The adult pair bond dissolved at the same time as the parent-
young relationship. The parents were no longer observed together at the
nest, which by this time had disintegrated and fallen through the mangrove.

Commonly 1 nestling ( less often 2 ) died in the nest. These were generally
the youngest and therefore the smallest. Because they were at a physical dis-
advantage in competing with older, more aggressive siblings for the limited
amount of food, they became emaciated and starved to death. On Grand Is-
land during 1973, I recorded a mortality rate from egg laying to the 2-week-
old age of 22.5%. Of the 34 nests, 1 contained one nestling, 15 two nestlings,
17 three nestlings, and 1 had four nestlings. Jenni (1969 ) calculated a mor-
tality rate of 35.8% for 28 nests for the same nestling period. Teal ( 1965 )
reported a mortality rate of 68% from egg laying until the young were
fledged. I attribute most of the nestling mortality on Grand Island to starva-
tion of the younger nestlings. I saw no evidence of predation by raccoons
\Procyon lotor): Boat-tailed Crackles (Quiscahis major) destroyed some
eggs in a small region of the heronry outside my study area. Nestlings oc-
casionally died after becoming entangled in mangrove branches during wind
storms or while wandering from the nest. A factor that may have contributed
to the low mortality rate on Grand Island is that the Barataria Bay estuarine
marsh is rich faunistically and provides a good nutrient source throughout
the breeding season ( Day et al. 1973 ) .

SUMMARY

Observations on the l)reeding behavior of the Louisiana Heron were made in a large
heronry on Grand Island, Barataria Bay, Plaquemines Parish, Louisiana. Male Louisiana
Herons with breeding soft-part colors set up large territories initially. The male ap-
peared tense at first but soon began performing Snap-Stretch and Circle Flight displays.
Soon after establishing a core display site, a male constructed the foundation of the
nest there. During the early courtship period the male defended his territory from both
intruding males and females. As additional males entered the breeding cycle and females
began to move through the heronry, the size of the male’s territory shrank.

At first a male threatened and chased away all females, but a persistent female was
soon tolerated on the edge cf the territory. The freciuency of Snap-Stretch and Circle
flight displays increased with time and in the presence of the female. With time, she
was allowed closer and closer to the nest. Assuming the Withdrawn Crouch, the female
was finally allowed onto the nest where copulation took place within a short time. Later
the female was allowed to occupy the center of the nest and began final construction.

rilK W ILSON lUILLKTIN • VoL 90, No. I, March 1978

nil

'riic male hroii^lit twifis to tlu- ffinale on the nest as both engaged in tlie (Greeting dis-
play and l>ill-nil)l)ling.

dlu* first eggs appeared 4 6 days after the initial copulation. On Cirand Island the
average clutch size was 3.0 ( range rr: 2-.3. mode = 3, N=99). Both sexes incubated.
When one mate returned, the pair engaged in the (Greeting display before the heron that
was relieved fh'w off.

Both sexes fed the young, at fiist r(*gurgitating small fish onto the floor of the nest.
Nestlings 2 weeks and older fed by grasping the hill of the parent. Adults brought food
to the brood 4-5 times a day. Many of the youngest nestlings died of starvation; chicks
of all ages occasionally died in accidents. A mortality rate of 22.4% was recorded for
nestlings on Grand Island for the period from egg laying through the second week after
hatching. The parent-young bond l)cgan to break down when the juveniles were 6-7
weeks old.

ACKNOWLEDGMENTS

This paper represents part of a thesis submitted in partial fulfillment of the require-
ments for the M..*^. degree in the Department of Zoology and Physiology. Louisiana State
I'niversity, Baton Rouge, Louisiana. I would like to thank George H. Lowery, Jr. of the
Museum of Zoology for providing both assistance and direction during my graduate re-
search. I am indebted to the Louisiana WildLife and Fisheries for furnishing me with
e(|uipment and use of the Marine Biological Laboratory on Grand Terre. The Louisiana
Ornithological Society awarded me a grant that helped defray the cost of research during
1973. Finally, I would like to thank my wife Linda for her assistance throughout the
entire investigation.

LITERATURE CITED

Bent, A. C. 1926. Life histories of North American marsh birds. Bull. L.S. Natl. Mus.
135:1-490.

Blaker, I). 1969. Behaviour of the Cattle Egret Ardeola ibis. Ostrich 40:75-129.

C(JTTRiLLE, W. P., AND B. I). CoTTRiLLE. 1958. Great Blue Heron: behavior at the nest.

Misc. Publ. Mus. Zook, Univ. Michigan, no. 102.

Day, J. W., W. G. Smith, P. R. Wagner, and W. C. Stowe. 1973. Community struc-
ture and carbon budget of a salt marsh and shallow bay estuarine system in Louisi-
ana. Center for Wetlands Resources, Louisiana State University, LSU-SG-72-04.
Huxley. J. 1922. Preferential mating in birds with similar coloration in both sexes.
Br. Birds 16:99-101.

Jenni, I). A. 1969. A study of the ecology of four species of herons during the breed-
ing season at Lake Alice, Alachua County, Florida. Ecol. Monogr. 39:245-270.
-Meani.ey, B. 1955. A nesting study of the Little Blue Heron in eastern Arkansas.
\\ ilson Bull. 67 :85-99.

Meyerriecks, a. j. 1960. Comparative breeding behavior of four species of North
American herons. Publ. Nuttall Ornithol. Club, no. 2.

. 1962. In Handbook of North American birds, Vol. I (R. S. Palmer, ed.). Yale

Univ. Press, New Haven, Conn.

Palmer, R. .S., (ed). 1962. Handbook of North American birds, Vol, 1. Yale Univ.

Press, New Haven, Conn.

Pratt, H. M. 1970. Breeding ecology of (ireat Blue Herons and Common Egrets in
central ('alifornia. Condor 72:407-416.

Rodgers • LOUISIANA HERON BEHAVIOR

Ridgway, R. 1912. Color standards and color nomenclature. Published by the author,
Washington, U.C.

Rodgers, J. A., Jr. 1977. Breeding displays of the Louisiana Heron. W'ilson Bull. 89:
266-285.

Siegfried, W. R. 1972. I'ood reciuirements and growth of Cattle Egrets in South Africa.
Living Bird 11:193-206.

Skead, C. J. 1966. A study of the Cattle Egret, Ardeola ibis, Linnaeus. Proc. Second
Pan-African Ornithol. Congr., Ostrich Suppl. 6:109-139.

Teal, J. M. 1965. Nesting success of egrets and herons in Georgia. Wilson Bull. 77:
257-263.

Thorpe, W. H. 1963. Learning and instinct in animals. Methuen, London.

Weber, W. J. 1975. Notes on Cattle Egret breeding. Auk 92:111-117.

MUSEUM OF ZOOLOGY, LOUISIANA STATE UNIV., BATON ROUGE 70803. PRESENT
ADDRESS: DEPT. OF BIOLOGY, UNIV. OF SOUTH FLORIDA, TAMPA 33620.

ACCEPTED 20 SEPT. 1976.

STATl'S AND NIJMKKICAL FLUCTUATIONS OF SOME NORTH
AMFIHCAN WADERS AI.ONG THE SURINAM COAST

A kip: L. Spa ans

'i'hroufihoul the year, the muddy coast of Surinam (South America) forms
a favorite haunt for large numbers of North American shorebirds (Haver-
sclimidt 1955, 1968). This paper deals with the status, numerical fluctuations,
and habitat selection of species that visit the Surinam coast regularly. Data
were gathered from April 1970 through May 1973.

STUDY AREA AND METHODS

Study area.- The Surinam coast is situated on the northeastern fringe of South America
at about 6° N latitude and between 54° and 57° W longitude (Fig. 1). Geographically,
it forms a part of the Guiana coast, the nearly 2000 km of muddy and sandy shore of
the Atlantic Ocean between the mouths of the Amazon and Orinoco rivers.

The coast of Surinam consists largely of vast tidal mud flats bordered on the higher
paits by forests of black mangrove i Avicennia germinans) . The flats alternate in space
and time from an accretion to an erosion coast; the succession of accretion and erosion
has a cyclic character. For a relatively small area along the Guyana coast, Diephuis
(1966) established that such a cycle takes about 30 years. The rapid succession of accre-
tion and erosion has resulted in a rather unstable shoreline. In a few places, the coast
is fringed with a narrow sandy beach. In 1971, 66% of Surinam’s 350 km shoreline was
in accretion, 24% w^as in erosion, 4% was stationary, while 6% w^as fringed with a sandy
beach ( P. A. Teunissen, pers. comm.).

The mud deposited along the Surinam coast originates from the Amazon. This river
yearly discharges large amounts of fine sediments into the Atlantic Ocean, which are
transported along the Guiana coasts by the Guiana Current. There the silt is deposited as
a w^atery sediment, called “sling mud” (l)iej)huis 1966). These depositions result in the
development of the tidal flats mentioned above. During exposure, the flats contain much
water, and as a result, are very soft. In general, it is impossible for men to walk on them
without sinking in. The flats are very rich in tanaids (Tanaidacea, Crustacea) and, during
exposure, constitute a very important feeding habitat for North American waders.

Along the erosion parts of the coast, the littoral zone consists mostly of a narrow, firm,
and tough hank of clay layers eroding from older deposits. Like the narrow' sandy beaches,
these clay hanks are less attractive for waders, having a much lower bird density than the
tidal flats.

Landward, the coastal fringe is bordered by a wide zone of shallow lagoons and of
brackish herbaceous swamps, broken by several low and narrow, wooded sand or shell
ridges lying parallel to the coast. The large complexes of lagoons also form an important
feeding habitat for waders. The lagoons are former mangrove forests in which the
Aiicennia has died in situ after prolonged inundation by sea water. For many years, the
trunks of the dead mangroves are a dominant feature of the lagoons. After some time, the
hare mud bottom of the lagoons may he covered by an herbaceous vegetation of halophytes,
predominately saltw<»rt iBatis marilima) and sea purslane (Sesuvium portiilacastnim) .
.Some lagoons have an extensive underwater vegetation of wigeon grass ^ Ruppia maritima) .

()0

Spaans • WADERS ALONG SURINAM COAST

Fig. 1. Map of the coastal area of Surinam showing the names of places mentioned in
the text.

The water level in the lagoons and the feeding possibilities for waders are highly variable
depending on the amount of precipitation, the freciuency of inundation by the sea,
and the amount of evaporation. The brackish herbaceous swamps found behind the belt
of mangroves are mainly covered by the spike rush i Eleocharis mutata) and other
Cyperaceae. During most of the year, the water level in these swamps is too high for
waders. In the long dry season, however, there are extensive shallow and muddy areas,
which attract large numbers of shorehirds.

Climate. — Surinam has a tropical climate; throughout the year, the mean daily tempera-
ture remains between 26° and 28°C. The amount of rainfall varies seasonally. Most rain
falls in April to July (long rainy season) with the least rain in September to November
(long dry season). Between November and April, there is a short rainy season (December
and January) and a short dry season (February and March), both with a moderate mean
monthly precipitation ( Meteorologische Dienst 1965). There is, however, much variation
in the onset of the seasons, both between coastal and inland Iccalities, and between years.
Table 1 shows the distribution of the monthly rainfall at the mouth of the Matapicakanaal
for 1961-70 and 1970-73.

Data collection. — To obtain a picture of the fluctuations in bird numbers, 1 made 6
series of regular counts:

1. From April 1971 through April 1973, the numbers of birds present during low tide
at a mud flat 10 km northwest of Paramaribo, locally known as “Weg naar Zee” (= road
to the sea), were counted at approximately 2 week intervals for one or more days from a
fixed point near the high tide water mark. The area surveyed had the shape of an
isosceles triangle with the observer placed at the apex ( 18° ) on the shoreline while the low
tide water mark formed the base. At low tide, the flat extended about 1 km seaward ( i.e.
height of the isosceles triangle).

THE WILSON BULLETIN • Vol. 90, No. 1, March 1978

Table 1

Monthly Rainfall (mm) at Matapicakanaal

Jan.

Kvhr.

.March

April

-May

June

July

Aug.

Sept.

Oct.

Nov.

Dec.

1970

n.a.*

313

168

244

164

225

1971

252

179

134

130

5.50

149

161

n.a.

1972

260

598

634

542

149

278

1973

197

1961-70

182

100

138

252

225

131

169

( means)

* n.a. — not available.

Most counts were made late in the afternoon, when the sun interfered the least witli
observation. Occasionally, however, I made counts in the early morning. Counts were
made with a 40 X 00 telescope. On most days, I made several counts hetw'een 1 h before
and 1 h after low' tide, from which I calculated the average numbers of birds present that
day. These average nundjers formed the starting point for further analyses.

2 and 3. From March 1971 through August 1972, I made counts at 2-4 week intervals
in 2 lagoons near Krofajapasi, and from May 1971 through November 1972 in 2 lagoons
near Motkreek. Both complexes of lagoons were situated just behind a sandy beach and
received sea water during spring tides. On several counting-days during the long dry
season, the lagoons near Krofajapasi were completely dry', with the result that many
w'aders had gone to feed elsewhere. In contrast, on some days during the long rainy
season, the water level was too high for most species of waders. The lagoons near
Motkreek also had a highly variable water level, but since these lagoons w'ere never
completely dry, the fluctuations in bird numbers during the long dry' season were less
than in the lagoons near Krofajapasi. In the long rainy season, however, there were some
counting-days with such a high water level that it must have had an unfavorable
effect on the numbers of waders present.

The length of the route taken in the lagoons of Krofajapasi covered about .7 km, that
in the lagoons near Motkreek about 3.5 km. The counts near Krofajapasi were made
in the afternoon, those near Motkreek in the morning. Along both transects, waders
were counted on both sides as far as they could be identified with certainty using 10 X 40
binoculars. At Krofajapasi, I also used a 40 X 60 telescope. .\s a result, the width of the
area covered was not the same for each species, and thus interspecific comparisons cculd
not he made.

4. From December 1970 through December 1971, at 2-4 week intervals, I counted
waders around low tide along the nearly 8 km sandy beach east of Krofajapasi, locally
known as “Bigi Santi'’ (= large beach). The counts were made between 06:00 and 09:00
(local time). This census included both the littoral zone, which was only some tens of
meters wide, and the beach. The latter was mainly covered by ipomoea (Ipomoea pes-
caprae) and sea bean {Canavalia maritirna). During the long dry season, patches of the
vegetation were burned. During the entire census period, the eastern end of the beach
was fringed on the seaward side with a narrow', firm and tough eroding clay hank. During
the first 9 months of the counting period, this hank was only a few hundred meters long,
hut after .Vugust 1971 it rapiilly increased to 1.5 km. Since the bird density on the clay

Spaans • WADERS ALONG SURINAM COAST

bank differed from that on the sandy beach, the numbers of birds counted in these
habitats will he given separately from September 1971 onward.

5. From March 1971 through October 1972, at Krofajapasi at 2-4 week intervals, I
counted the numbers of Spotted Sandpipers (Actitis macularia) flying down the creek to
their nightly roosts outside the mangrove swamps. Counts were made from about 17 :00
until dark, which occurred between 18:20 and 19:00 depending on the time of the year.

6. From December 1970 through December 1971, Spotted Sandpipers were also counted
along the nearly 6 km long Matapicakanaal. This canal runs through a former plantation
area, now mainly covered by mangroves. The counts were made at various times of the
day, and at various tide levels.

Presentation of the results. — The results of the censuses are given as mean numbers per
counting-day in 10 or 15 day periods. The status of each species will he given by an
estimate of the maximum number present at one day during the season (s) involved.
These estimates have been obtained by extrapolation of the transect counts for the entire
coast, adjusted by the percentages of suitable habitat. Considering the small number of
transects counted and the small area of each habitat they covered, these estimates, of
course, give only a very rough picture of the numbers of birds. As an index of abundance,
the following scale has been used:

very large numbers —100,000 or more individuals
large numbers — 50,000-100,000 individuals

rather large numbers — 10,000-50,000 individuals
rather small numbers — 2500-10,000 individuals
small numbers — 500-2500 individuals

very small numbers — 1-500 individuals.

Data on the status only refer to the numbers present in the coastal area landward up to
and including the zone of brackish herbaceous swamps. For some species, therefore, the
number of birds in .‘Surinam may he higher than the numbers mentioned here.

RESULTS

In the following list, an account is given of the occurrence of the species
which visit the Surinam coast yearly. For each species, data on the status,
habitat selection, and numerical fluctuations will be given:

Black-l>elliecl Plover iPluvialis squatarola) . — This species is present
throughout the year in rather small numbers, possibly in rather large numbers.
It can he observed everywhere along the coast, both on soft or hard mud, and
on sandy substrates.

Numbers of any importance were counted in the transects of Bigi Santi
and Krofajapasi only. In both transects, the species was present throughout
the year. In neither of the transects, however, was a clear-cut seasonal trend
observed.

American Golden Plover {Pluvialis dominica) . — This species is a transient
in very small numbers from September through November. It is mainly a
bird of dry inland areas. Along the coast, I observed the species on the sandy
beach of Bigi Santi in an area where ihe vegetation had recently been burned,

THK WILSON iaJLIi:riN • VuL 90, No. 1, March 1978

SEMIPALMATED PLOVER

MONTHS

Fig. 2. Seasonal variations in numbers of Seinipalinated Plovers along the Surinam
coast. A. Beach of Bigi Santi, January through December 1971. The numbers counted on
the sandy beach are indicated by cross-hatching; those on the clay hank are unshaded.
B. Lagoons near Krofajapasi, March 1971 through April 1972. A cross indicates that
no water was in the lagoons. The numbers at the top of the columns show the numbers
of counts on which each datum is based.

and in parts of the lagoons near Motkreek that had dried up, between 15 Sep-
tember (1 bird) and 17 November (3 birds). My observation dates match
rather well the period of fall migration mentioned by Haverschmidt (1969).
L ntil now, the latest date for a fall migrant was 8 Nov'ember (Haverschmidt
op. cit. ) .

Seiiiipaliiiated Plover [Charadrius semipalmatus) . — This species is present
throughout the year in rather large numbers. It is mainly distributed along
the coast where it shows a strong preference for tidal flats and muddy lagoons.

Data on numerical fluctuations are available for the beach of Bigi Santi
and for the lagoons near Krofajapasi (Fig. 2). At Bigi Santi, the species was
seen throughout the year, except for a period of 2 months from mid-June
through August. A pronounced peak in numbers occurred during late fall and
early winter. In the lagoons near Krofajapasi, the species was observed

Spaans • WADERS ALONG SURINAM COAST

throughout the year with the same peak during the fall and winter periods.
After December, numbers fell rapidly, both at Bigi Santi and near Krofajapasi.
Since the drop in numbers coincides with the onset of the short rainy season,
and hence with an enlargement of feeding areas in the lagoons, I suggest that
it reflects a dispersal of birds to other feeding areas rather than a departure
from Surinam.

During the spring, numbers remained low with no peak representing the
passing of spring migrants. However, a small increase in numbers occurred
in June. Since tbe increase was also noticed in other lagoons of the Krofajapasi
area, I suggest that it reflects the arrival of summer visitors, probably from the
South, rather than a concentration of birds that did not migrate to the North.

Upland Sandpiper { Bartramia longicaiida). — Along the coast, this species
is a transient in very small numbers in September and an irregular winter
visitor. There, I observed it mostly on the sandy beach where the vegetation
was rather open or where it had recently been burned. My sightings occurred
between 1 September (2 birds) and 19 February (1 bird). Observation dates
all fall within the period mentioned by Haverschmidt (1966).

Whimhrel {Numenius phaeopus). — This species is present throughout the
year in rather small numbers with highest numbers from August through
October. It is confined to muddy substrates along the coast.

None of the transect counts showed significant numbers. The statement on
status mentioned above is based on (jualitative data from outside the transects.

Lesser Yellowlegs {Tringa flavipes). — This species is a transient and winter
visitor in very large numbers from mid-July through early May and a summer
visitor in small to rather small numbers. The species is one of the most
abundant waders on the tidal flats and in the shallow lagoons and brackish
herbaceous swamps. Further inland, it is also numerous on flooded ricefields.

Data on numerical fluctuations are available for the mud flat near Weg
naar Zee and for the lagoons near Motkreek and Krofajapasi (Fig. 3).
Near Weg naar Zee, numbers increased from late July through mid-August,
after which they remained high until early April. Near Motkreek, an increase
in numbers occurred from mid- July through early August, after which they
remained high until early March. Numbers then decreased rapidly. Near
Krofajapasi, a moderate increase in numbers was seen in mid- July, followed
by a decrease in September and an absence until mid- December. This was
probably a result of poor feeding conditions there. From December through
February, numbers remained moderate, followed by a peak abundance during
March.

It might appear from the above data that a mass arrival of southbound
transients did not occur before August. This, however, is not true. Large

THE W lESON BULLETIN • VuL 90, No. 1, March 1978

LT)

Cxi

MONTHS

Fig. 3. Seasonal variations in numbers of Lesser Yellowlegs along the Surinam coast.
A. Mud flat near eg naar Zee, April 1971 through April 1973. B. Lagoons near Motkreek,
October 1971 through November 1972. C. Lagcons near Krofajapasi, March 1971 through
August 1972. An arrow indicates a very high water level in the lagoons. P indicates a
partial count. Other conventions as in Fig. 2.

numbers arrive in July but evidently, few make a stop for any length of time.
This was corroborated by observations of large numbers of Lesser Yellowlegs
beading eastward along the coast in the second half of that month.

On the mud flat near Weg naar Zee the highest numbers occurred from
August through December ( long dry season I and in the lagoons near Motkreek

Spaans • WADERS ALONG SURINAM COAST

and Krofajapasi from December through March (short rainy season and
short dry season). This indicates a shift from the mud flats to the lagoons
with the onset of the rainy season.

The drop in numbers occurring after March near Motkreek and Krofajapasi
suggests a mass departure from March onward. This was corroborated by
observations of Lesser Yellowlegs departing the Motkreek lagoons on 10 March
1973. Data from mid-May through early July indicate that the species is a
regular but not numerous summer visitor.

Greater Yellowlegs {Tringa melanoleuca) . — This species is a transient in
large numbers and a winter visitor in rather large numbers from mid-Septem-
ber through early May; it is a summer visitor in rather small numbers. Like
the former, this species shows a strong preference for tidal flats and shallow
lagoons; further inland, it is also common on flooded ricefields.

Data on numerical fluctuations are available for the mud flat near Weg
naar Zee and for the lagoons near Motkreek (Fig. 4). The species was seen
throughout the year in both transects with peak abundances during September—
October (Weg naar Zee ) , in March ( Weg naar Zee ) , and in April ( Motkreek j .
The absence of a peak near Motkreek during the fall is somewhat puzzling.
The peaks in fall and spring may coincide with the passage of southbound
and northbound transients.

Solitary Sandpiper [Tringa solitaria). — This species is a transient and
winter visitor in rather small numbers from late July through early May; it is
possibly present in rather large numbers during the fall. Along the coast, this
is mainly a bird of shallow lagoons and brackish herbaceous swamps; further
inland, it is also a common visitor of freshwater pools and ditches, and of
flooded ricefields.

I observed the species during the transect counts only occasionally. The
statement on status mentioned above is based on qualitative data from outside
the transects. My observations of a Solitary Sandpiper on 19 July 1972 and
one on 9 May 1971 are tbe earliest and latest dates respectively, for this species
in Surinam ( see Haverschmidt 1968 ).

Spotted Sandpiper [Actitis macularia). — This species is a transient and
winter visitor in rather large numbers from early July through early June;
it is a summer visitor in very small, local numbers. The species is not con-
fined to the coast and may be found everywhere that water is present. Along
the coast, it shows a preference for shallow and muddy lagoons, but it may
also be commonly found along creeks and canals, on firm and tough clay banks
emerging from eroding coastline, and on higher parts of mud flats. During
fall migration, it is also numerous on sandy beaches.

(){\

I HK W ILSON BULLETIN • VoL 90, No. 1, March 1978

MONTHS

Fig. 4. Seasonal variations in numbers of Greater Yellowlegs along the Surinam coast.
A. Mud flat near W'eg naar Zee, April 1971 through April 1973. B. Lagoons near Mot-
kreek, October 1971 through November 1972. Conventions as in Fig. 2 and 3.

Data on numerical fluctuations are available for Krofajapasi, Matapicaka-
naal and the beach of Bigi Santi (Fig. 5). The first fall migrants arrive
(luring early July. The earliest dates, based on records from places where
the species was absent in the weeks before, are: 2 July 1971 (2 birds), 5
July 1972 (8 birds; these birds must have arrived during the night since
the species was still absent the day before at each place of observation), and
6 July 1970 (1 bird). After then numbers built up rapidly, reflecting a mass
arrival of migrants. From August through May, the numbers of birds flying
down Krofajapasi Creek fluctuated heavily; highest numbers were in January
and lowest numbers were in September. Along the Matapicakanaal, the num-
bers remained high through January with much lower numbers occurring dur-

Spaans • WADERS ALONG SURINAM COAST

SPOTTED SANDPIPER

0-^ 1 ^ ^ W . I ,

J FMAMJJASOND
MONTHS

Fig. 5. Seasonal variations in numbers of Spotted Sandpipers along the Surinam coast.
A. Krofajapasi Creek, March 1971 through October 1972. B. Matapicakanaal, December
1970 through December 1971. C. Beach of Bigi Santi, December 1970 through December
1971. Conventions as in Fig. 2.

ing the spring months. At Bigi Santi, numbers decreased sharply during
September, indicating that the peak of fall transients passed through mainly
in August and early September. During the spring months, no migration peak
was observed in any of the transects.

During June, only few Spotted Sandpipers were seen, either in or outside
the transects. Of my 14 observations (25 birds) in June, 8 (18 birds) were

THK WILSON BULLETIN • VoL 90, No. I, March 1978

WILLET

< ■ 2 ^B2±1 ^ 1 J

^ j 1 ■■ 1 1 1 1

F M A M J J A S O
MONTHS

N D

Fig. 6. Seasonal variations in numbers of Vi illets along the Surinam coast ( mud flat
near \\ eg naar Zee, May 1971 through April 1973). Conventions as in Fig. 2.

the entire month. The other 6 records ( 7 birds) were from other places, and
all fell in the first week of June. The latter, therefore, are likely to be late
spring migrants rather than summer visitors. The data show, however, that
near Krofajapasi Spotted Sandpipers were present during the whole month of
June. I feel justified in considering these as summering birds. Only 2 June
records have previously been reported ( Haverschmidt 19681 .

Willel (Catoptrophorus semipalmatus) . — This species is a transient in rather
large numbers, possibly in large numbers, from early July through mid-
August; it is present during the other months in rather small numbers. The
species is confined to the coast where it shows a strong preference for tidal
flats. During the fall migration, however, large flocks may also be encountered
in lagoons that have dried up.

Data on numerical fluctuations are only available for the mud flat near
Weg naar Zee ( Fig. 6 ). The species was seen here throughout the year with a
peak abundance from late July through mid-August, after which the mean
numbers fluctuated around a low level. In March (1973) and April (1972),
the numbers counted were somewhat higher, perhaps reflecting the passage of
small numbers of spring migrants.

From these data, it might appear that a mass arrival of southbound tran-
sients did not occur before the end of July, hut this is not true. Elsewhere along
the coast, the species was already numerous in mid-July. In 1972, I observed
several flocks of 10-50 birds heading east at Eilanti as early as 4 July. Al-
though peak numhers are over by late August, transients may pass until well
into October.

of birds flying down Krofajapasi Creek during the routine counts throughout

Spaans • WADERS ALONG SURINAM COAST

SHORT- BILLED DOWITCHER

MONTHS

Fig. 7. Seasonal variations in numbers of Short-l)illed Dowitchers along the Surinam
coast. A. Mud flat near Weg naar Zee, April 1971 through April 1973. B. Lagoons near
Motkreek, September 1971 through November 1972. Conventions as in Figs. 2 and 3.

Ruddy Turnstone [Arenaria interpres ) . — This species is present throughout
the year in rather large numbers. It can be found everywhere along the
coastal shore where the substrate is firm.

Counts were made along the beach of Bigi Santi and in the lagoons near
Krofajapasi. The numbers counted in these transects show no clear-cut
seasonal trend, except for an increase during the fall and early winter on the
elay bank at the eastern end of Bigi Santi coincident with an increase in the
length of this hank. In neither of the two transects was there any indication of
a migration peak in the northern spring. In both transects, rather fair numbers
were present during the northern summer months.

TllK WILSON lUILLKTIN • Vul. 90, No. 1, March 1978

Fig. 8. Seasonal variations in numliers of Red Knots Spooled data) along the Surinam
coast in 1970-73. Conventions as in Fig. 2.

Short-hilled Dowiteher ( Limnodromus griseus) . — This species is a transient
in very large numbers during the fall, from mid-August through early October,
and during the spring, from early March through late May; it is a winter and
summer visitor in small numbers. This species was found almost exclusively
in lagoons and on tidal flats. On the latter, it showed a strong preference for
the zone of sling mud near the low-water mark.

Data on numerical fluctuations are available for the mud flat near Weg
naar Zee and for the lagoons near Motkreek (Fig. 7). Near Weg naar Zee,
numbers were high from late summer through early fall and in the spring.
Numbers were low during the northern winter, and the species was absent
during the northern summer. Near Motkreek, highest numbers were seen in
late summer and early fall, and lowest numbers were seen during the northern
winter and summer. The low spring numbers probably resulted from high
water levels during the long rainy season.

Red Knot \Calidris canutus) . — This species is a transient in small numbers
from mid- August through late October and in early May; it is a summer and
winter visitor in very small numbers. Most Red Knots were observed on
firm and tough clay banks emerging from eroding coastline and in shallow
lagoons. I never observed the species on the soft tidal flats.

Since the species was not observed freiiuently in any of the transects all
observations of knots along the coast have been lumped to obtain an idea about
its occurrence ( iMg. 8 l. 1 he species was seen throughout the year with peaks
from August tb rough October, and in May.

Spaans • WADERS ALONG SURINAM COAST

Fig. 9. Seasonal variations in numbers of Sanderlings along the Surinam coast
(beach of Bigi Santi, January through December 1971). Conventions as in Fig. 2.

Sanderling ( Calidris alba I . — This species is a transient and winter visitor
in small, possibly in rather small numbers, with highest numbers in fall; it is
a summer visitor in very small numbers. The species, confined to the coast, was
almost exclusively found on sandy beaches and on firm and tough clay banks
emerging from eroding coastline.

Data on numerical fluctuations are available only for the beach of Bigi
Santi (Fig. 9). The species was seen throughout the year with highest num-
bers from September through December. On an average, the lowest numbers
were seen from May through August, but in one census ( 17 June 1971) 33
birds were counted — almost as many as the average number during the first
three months of the year. In September, there was an increase in numbers on
the sandy beach. Some tens of birds were also observed then on the clay
hank fringing the shore at the eastern end of the beach, where the species had
not been seen in the months before. On the sandy beach, numbers remained
more or less constant throughout the rest of fall, except for a peak in the second
half of September. However, on the clay hank, parallel to an increase in its
length, numbers increased from October onward. A pronounced migration
peak in spring was not observed, hut it should he noted that no data were
available for April.

TIIK WII.SUN ItUl.I.KTIN • VoL VO, No. I, .March I97R

SEMIPALMATED SANDPIPER

MONTHS

Fig. 10. Seasonal variations in numbers of Semipalmated Sandpipers along the Suri-
nam coast. A. Mud flat near Weg naar Zee, April 1971 through April 1973. B. Lagoons
near Krofajapasi, March 1971 through April 1972. Conventions as in Fig. 2.

S(‘nii|)aliiiat(Ml Sandpiper iCalidris pusilla) . — This species is a transient
and winter visitor in very large numbers from mid-August through mid-June,
and a summer visitor in rather large numbers, possibly in large numbers. The
species is mainly confined to the coast, where it shows a strong preference
for tidal flats and shallow lagoons.

Data on numerical fluctuations are available for the mud flat near Weg
naar Zee and for the lagoons near Krofajapasi (Fig. 10). Numbers near

Spaans • WADERS ALONG SURINAM COAST

Weg naar Zee increased gradually starting about mid-August with peak
abundance in early November, after which they decreased. During the north-
ern winter, numbers fluctuated heavily with fewer present in January and
February than in December. From late February onward, numbers were
somewhat higher, with a small peak in early May. After this, numbers de-
creased to the low summer level.

Near Krofajapasi, no increase in numbers was observed after the onset of
the fall migration. This was probably due to poor feeding possibilities there
during the long dry season. After the onset of rains in January 1972, num-
bers were higher with peak abundance in April. Relatively high numbers
were also present in June and July, when numbers near Weg naar Zee were at
their lowest. This suggests that summering birds may concentrate at certain
localities, which was corroborated by observations of large numbers elsewhere
along the coast. For example, 30 June 1971, at least 12,000 Semipalmated
Sandpipers were present on the tidal flats east of the mouth of the Coppename
River, while in early July 1972, several thousand were present on the flats near
Eilanti.

Western Sandpiper {Calidris mauri) . — This species is a winter visitor in
small numbers, possibly in rather small numbers, and probably a summer
visitor in very small numbers. The Western Sandpiper is confined to the coast,
where it frequents tidal flats and shallow lagoons. In western Surinam, they
are probably more numerous than near Paramaribo. During 1970-73, around
high tides, we caught only 2 C. mauri against nearly 3700 C. pusilla in the
mangrove swamps near Weg naar Zee. In Nickerie, in the west of the country,
however, Mr. W. E. van der Schot (pers. comm.) found at least 8 C. mauri
(of which several are now in the Zoological Museum, Amsterdam ) among 54
“peeps” shot by a hunter on 12 November 1972.

Least Sandpiper [Calidris minutilla) . — This species is a transient and
winter visitor in large numbers from mid-July through mid- June, and a sum-
mer visitor in small, local numbers. Along the coast, the species shows a prefer-
ence for muddy lagoons and brackish herbaceous swamps. As far as I know, it
avoids the exposed areas of the tidal flats but is rather numerous on open sites
in the mangrove forests. The species is not restricted to the coast; further in-
land, it is a common species along ditches, in freshwater swamps, and in
flooded ricefields.

Eig. 11 shows the numerical fluctuations in the lagoons near Motkreek and
Krofajapasi. Near Motkreek, numbers increased during mid- July. From early
August through early May, the species was present in fluctuating numbers
with peak abundances in mid-August, early February, and early May. From
late May through late June, the species was not observed in this transect.

MEAN NUMBERS

rHK WILSON BULLETIN • Vof. <)0, No. I, March 1978

J F M A M

140-1

120-

100- ,

80- I

LEAST SANDPIPER

60-

J FMAMJ JASOND
MONTHS

Fig. 11. .Seasonal variations in nunihers of Least Sandpipers along the Surinam coast.
A. Lagoons near Motkreek, May 1971 througli Noveinher 1972. B, Lagoons near Krofa-
japasi, March 1971 through August 1972. Conventions as in Figs. 2 and 3.

Spaans • WADERS ALONG SURINAM COAST

Near Krofajapasi, numbers increased gradually from mid- July through Sep-
tember, after which numbers fluctuated heavily with peaks in mid-November
and early April. After early May, numbers decreased, but the species was
present through mid-June. In late June and early July, no Least Sandpipers
were observed in this transect.

Since the peak abundances were found in different periods of the year
for the two complexes of lagoons, I suggest that they reflect fluctuations in
feeding opportunities in the lagoons rather than differences in abundance as
a result of arrival or departure of transient birds. In both transects, no birds
were observed between departure of the last spring migrants and arrival of
the first fall migrants. In this period, however, the species was observed
regularly near Krofajapasi outside the transects. In the second half of June
and first half of July 1971 and 1972, on five different dates, a total of at least
74 Least Sandpipers was observed. From these observations it may be con-
cluded that at least locally, the species stays over during the northern summer.
Summer records have not been reported previously ( Haverschmidt 1968, pers.
comm, j .

White-rumped Sandpiper (Calidris juscicollis) . — This species is a tran-
sient in rather large numbers from mid-August through late November and
in rather small numbers from early April through mid-June; it is probably an
irregular winter visitor in small or very small numbers. The species was ob-
served almost exclusively in lagoons and brackish herbaceous swamps. As
far as I know, it avoids the lower zones of tidal flats. It may occur oc-
casionally, however, in the higher zones of the littoral. The species also occurs
further inland along freshwater pools and ditches.

Fig. 12 shows the seasonal variations in numbers in the lagoons near
Motkreek and Krofajapasi. There was a mass arrival during late August
(earliest date, 20 August 1972) and large numbers were also encountered in
September. In October and November, however, only small numbers were
seen. Most fall transients therefore pass through Surinam during the last ten
days of August and in September. This is corroborated by a strong easterly
migration parallel to the coast during these months ( e.g. 31 August 1971,
when, between 09:30 and 10:30, 688 birds in 61 flocks were observed near
Motkreek flying low over the ground in an easterly direction). My latest fall
observation date for this species is 28 November 1971. During the spring
there was a low peak in May with the first birds arriving in early April
(earliest date, 8 April 1972 I ; the last birds departed mid-June (latest date,
21 June 1972).

Fall migration started much earlier and spring migration was extended
much longer than mentioned by Haverschmidt ( 1968 ) . Although I was on the

TIIK WILSON BULLETIN • Vol. 90, No. 1, March 1978

WHITE-RUMPED SANDPIPER

400 1

300

200

100-

LT)

Cxi ^
LU

CQ 0-

Z 100

^ 80-
60
40
20

1 ’ 1 ’ ip JJ. ’ 1 '

J ' M '~A M J " S O N D

1 1

° ' J "~E

11 ^

M A

2 2 12 1

M J J A
MONTHS

Fig. 12. Seasonal variations in numbers of White-rumpecl Sandpipers along the Surinam
coast. A. Lagoons near Motkreek, May 1971 through November 1972. B. Lagoons near
Krofajapasi, March 1971 tbrough August 1972. Conventions as in Figs. 2 and 3.

look-out for wintering and summering birds, I did not locate any. Mr. F.
Haverschmidt iin litt.) told me, however, that he observed the species several
times in the swamps near Maasstroom in December 1963 and February 1964,
suggesting that in some years the species may stay over during the northern
w inter.

Pectoral Sandf)i|)cr (Calidris melonotos ) .—In Surinam, this is more a bird
of iidand than of coastal habitats. Along the coast, I observed the species

Spaans • WADERS ALONG SURINAM COAST

STILT SANDPIPER

MONTHS

Fig. 13. Seasonal variations in numbers of Stilt Sandpipers along the Surinam coast.
A. Lagoons near Motkreek, Septemlier 1971 through November 1972. B. Lagoons near
Krofajapasi, March 1971 through August 1972. Conventions as in Figs. 2 and 3.

occasionally in shallow lagoons. It is probably more numerous in the brackish
herbaceous swamps, and as far as I know, it avoids the tidal flats. Because of
its scarce occurrence along the coast, I am not able to add new facts about its
stay in Surinam (see Haverschmidt 1968).

Stilt Sandpiper {Micropalama himantopus) . — This species is a fall tran-
sient in rather large numbers and a winter visitor in rather small numbers
from early August through mid-May; it is probably a summer visitor in small
or very small numbers. The species is confined to shallow lagoons and brackish
herbaceous swamps. I have no observations of this species from the tidal flats.

Fig. 13 shows the fluctuations in Stilt Sandpiper numbers in the lagoons

THE WILSON BULLETIN • VoL 90, No. 1, March 1978

near Motkreek and Krofajapasi. The first fall transients arrive during early
August. Ihe two transects respectively show peak abundance in mid-August
and late August, reflecting a mass influx of southbound transients during
this period. From September onward, numbers decreased in both complexes
of lagoons, albeit more rapidly in the lagoons near Krofajapasi than in those
near Motkreek. This difference possibly resulted from deteriorating feeding
conditions in the former. During the northern winter and spring, the species
was present in fluctuating numbers in both complexes of lagoons. No migra-
tion peak during the spring months was observed. The presence of large
numbers of Stilt Sandpipers in nuptial plumage in the first half of May 1970
and 1971, indicates that spring departure continues until mid-May. During
the northern breeding season Stilt Sandpipers were counted during early
June only near Krofajapasi.

COXCLUSIONS AND DISCUSSION

Occurrence. — Of the 24 species and subspecies of North American waders
known to visit the Surinam coast, 20 do so regularly. Most of these are
present in Surinam throughout the year I Haverschmidt 1955, and 1968; and
this paper). Exceptions are the American Golden Plover, Upland Sandpiper,
Solitary Sandpiper, White-rumped Sandpiper, and the Pectoral Sandpiper,
species for which no summer records are known. Records of the American
Golden Plover are also lacking from the northern winter.

Of the 20 regular visitors, the Lesser Yellowlegs, Short-billed Dowitcher
and the Semipalmated Sandpiper occurred during the study in very large
numbers (maximum numbers present at one day estimated at >100,000), the
Greater Yellowlegs, Least Sandpiper, and possibly the Willet in large numbers
(maximum numbers 50,000-100,000), and the Semipalmated Plover, Spotted
Sandpiper, Ruddy Turnstone, White-rumped Sandpiper, and the Stilt Sand-
piper in rather large numbers (maximum numbers 10,000-50,000). The
remaining species, except 2, occurred in maximum numbers lower than 10,000.
The exceptions are the Black-bellied Plover and the Solitary Sandpiper,
which were classified as occurring in rather small, possibly in rather large
numbers.

Fall migration routes. — For several species the occurrence and abundance
in fall differ to some extent from those in northeastern Venezuela as reported
by McNeil ( 1970) . In Surinam, peak numbers of the American Golden Plover,
Willet, Short-billed Dowitcher, Red Knot, and the V hite-rumped Sandpiper
are reached much earlier during the fall than they are in northeastern Vene-
zuela. For the Red Knot and the White-rumped Sandpiper, peak numbers in
Surinam are also much higher than in northeastern Venezuela. These data may
indicate that these species reach Surinam from North America mainly through

Spaans • WADERS ALONG SURINAM COAST

a non-stop flight across the Atlantic Ocean. There is much circumstantial evi-
dence that a non-stop flight across the Atlantic Ocean in late summer and fall is
a common feature among waders. Radar studies along the coast of the Canadi-
an Atlantic provinces (Richardson 1974), Massachusetts I Drury and Keith
1962, Nisbet 1963 ), and New Jersey ( Swinehroad 1964 ) have shown that large
numbers of North American waders on southbound migration move out to
sea in ESE-SSE directions, . . as if on a non-stop flight to northern South-
America” ( Drury and Keith 1962 ). In addition, McNeil and Cadieux ( 1972)
and Burton and McNeil 1 1975 ) working on southbound migrating waders at
the Magdalen Islands in the Gulf of St. Lawrence and at Sable Island off Nova
Scotia showed that many North American waders leave the Canadian Atlantic
provinces with enough energy reserves to fly non-stop over the Atlantic Ocean
to reach the Lesser Antilles and the northeastern coast of South America.
Moreover, the regular occurrence of various species of North American waders
at Bermuda in late summer and fall (Wingate in Drury and Keith 1962) also
points to a transoceanic flight to South America, as do the recoveries of birds
handed at the Magdalen Islands and at Sable Island (NlcNeil and Burton
1973, Burton and McNeil 1975 ).

The Greater Yellowlegs and the Semipalmated Sandpiper, on the other
hand, reach their fall peak much later in Surinam than in northeastern
Venezuela. Eor the Semipalmated Sandpiper, the Surinam peak follows a large
departure of adults from Venezuela (McNeil 1970), which might indicate a
relationship between counts in the two areas.

Spring migration routes. — In spring, too, some striking differences in occur-
rence and abundance were found between Surinam and northeastern Venezuela.
In Surinam, a large spring migration peak was observed in the Short-billed
Dowitcher and the Red Knot, and possibly also in the Greater Yellowlegs and
the Willet, whereas in northeastern Venezuela none of these species showed a
significant increase in numbers. In Venezuela, on the other hand, the Semi-
palmated Plover, Lesser Yellowlegs, Spotted Sandpiper, Ruddy Turnstone,
Semipalmated Sandpiper, W hite-rumped Sandpiper, Stilt Sandpiper, and pos-
sibly the Least Sandpiper did show a clear spring migration peak, whereas in
Surinam no migration peak, or only a very small one, was observed in these
species.

On the basis of these data, I suggest that the spring migration route of
Short-hilled Dowitchers and Red Knots, and possibly also of Greater Aellow-
legs and Willets, wintering south of the Guianas, is mainly to the Guiana coast,
and from there non-stop across the Atlantic Ocean to North America. The
latter suggestion is supported by observations of the departure of several
flocks of Short-hilled Dowitchers near Weg naar Zee in late April and early
May 1971 and 1972, in directions varying between 332° and 351° (Spaans,

THE WILSON BULLETIN • Vol. 90, No. 1, March 1978

uiipiil)!. data), dlie arrival of waders at Cape Cod, Massachusetts, in May from
southeastern directions ( Drury and Keith 1962, Nisbet 1963) also points to
some transoceanic flifjhts in spring.

Most other species, however, possibly take their spring migration route
mainly across the South American continent hy-passing the Guianas. Such a
route is not unimaginable since many waders follow a route through the West
Indies and the Mississippi Valley (Cooke 1912 I .

The reason for the differences in spring migration routes may lie with the
geographical areas where the birds breed: a transoceanic flight for birds
breeding in the most eastern parts of North America, and a route through the
Caribbean and the Mississippi Valley for birds breeding in more western areas.

SUMMARY

During 1970-73, regular counts of North American waders were made along the Surinam
coast to obtain quantitative data on their status and numerical fluctuations through the
year. An extrapolation of these counts for the entire coast renders it likely that maximum
numbers for the Lesser Yellowlegs, Short-l)illed Dowitcher, and the Semipalmated Sand-
piper may amount to over 100.000, for the Greater Yellowlegs, Least Sandpiper, and
I)ossibly the illet to 50-100,000. for the Semipalmated Plover, Spotted Sandpiper, Ruddy
Turnstone, hite-rumped Sandpiper, and the Stilt Sandpiper to 10-50,000. All other
species, except the Black-bellied Plover and the Solitary' Sandpiper, whose numbers may
possibly amount to over 10,000, are less numerous.

Comparison of the data with counts from northeastern \ enezuela suggests that in the
fall many species reach the coast of Surinam through a non-stop flight across the Atlantic
Ocean. In spring, many waders that spent the winter south of the Guianas seem to migrate
across the South American continent, by-passing the Guianas. Only the Short-billed
Dowitcher, the Red Knot, and pcssibly the Greater Yellowlegs and the W illet. mainly fly
directly to the Guianas and from there non-stop across the Atlantic Ocean to North
America.

ACKNOWLEDGMENTS

My thanks are due to Dr. J. P. .^chulz. Surinam Forest Service, who has called attention
to the need for ornithological research along the Surinam coast, to Professor M. F.
Mhrzer Bruijns, Department of Nature Conser\ation, Agricultural L niversity, W ageningen,
and Professor K. H. \ oous. Free Lniversity, Amsterdam, for supervision of the study. The
study was financed by the Netherlands Foundation for the Advancement of Tropieal
Research ( O I RO i , The Surinam Forest Service provided boats and manpower to
make possible the counts near Krofajapasi and surroundings, and to survey other rather
inaccessible parts of the eoast. My thanks are due to Mr. J. N. M. Verholt for drawing
the figures, to Mr. IL A. Reichart for correcting the English text, and to Dr. D. G. Ainley
and Dr. R. McNeil for their eritical review of the manuscript. This paper was written
while I held a temporary appointment at the Institute for Ecological Research. Arnhem.
The Netherlands.

LITERATURE CITED

Bi hton, j. a.M) R. McNkil. 19<5. Les routes de migration automnale de treize especes
d’oiseaux de rivage Nord-Americains. Rev. Geogr. Montr. 29:305-334.

Spaans • WADERS ALONG SURINAM COAST

Cooke, W. W'. 1912. Distribution and migration of North American shorebirds. U.S.

Dept. Agric. Biol. Surv., Bull. 35, revised.

Diephuis, J. G. H. R. 1966. The Guiana coast. Tijdschr. Kon. Ned. Aardr. Gen. 83:145-
152.

Drury, W. H. and J. A. Keith. 1962. Radar studies of songbird migration in coastal
New England. Ibis 104:449-489.

Haverschmidt, F. 1955. North American shore birds in Surinam. Condor 57:366-368.

— -. 1966. The migration and wintering of the Upland Plover in Surinam. W ilson

Bull. 78:319-320.

. 1968. Birds of Surinam. Oliver & Boyd, Edinburgh and London.

. 1969. The migration of the American Golden Plover through Surinam. ilson

Bull. 81:210-211.

McNeil. R. 1970. Hivernage et estivage d’oiseaux aquatiques Nord-Americains dans
le Nord-Est du Venezuela ( mue, accumulation de graisse, capacite de vol et routes
de migration). L’oiseau et R.F.O. 40:185-302.

AND J. Burton. 1973. Dispersal of some southbound migrating North American

shorebirds away from the Magdalen Islands, Gulf of St. Lawrence, and Sable Island,
Nova Scotia. Carih. J. Sci. 13:257-278.

AND F. Cadieux. 1972. Fat content and flight-range capabilities of some adult

spring and fall migrant North American shorebirds in relation to migration routes on
the Atlantic coast. Nat. Can. 99:589-605.

Meteorologisciie Dienst. 1965. Climatological tables various elements Paramaribo.
Period 1931-1960, monthly means. Serie 3. No. 2. Second edition.

Nisbet, I. C. T. 1963. Measurements with radar of the height of nocturnal migration
over Cape Cod, Massachusetts. Bird-Banding 34:57-67.

Richardson, W. J. 1974. Spring migration over Puerto Rico and the Western Atlantic,
a radar study. Ibis 116:172-193.

SwiNEBROAD, J. 1964. The radar view of bird migration. Living Bird 3:65-74.

SUEtLXAM FOREST SERVICE, PARAMARIBO, SURINAM, AND INSTITUTE FOR ECOLOGI-
CAL RESEARCH, ARNHEM, THE NETHERLANDS. ( MAILING ADDRESS: RESEARCH
INSTITUTE FOR xNATURE MANAGEMENT, KEMPERBERGER^VEG 67, ARNHEM,
THE NETHERLANDS). ACCEPTED 10 FEB. 1977.

FEEDING OE NESTLING AND ELEDGLING
EASTERN RLLEHIKDS

Hk.nkdict C. Flnkowski

Several workers ( Forl)es 1903, Beal 1915, Cottam and Knappen 1939,
Davison 19621 have described the prey consumed by adult Eastern Bluebirds
{Sialia sialis). No definitive studies, however, have been done on the diet of
nestling and fledgling bluebirds. In this paper I summarize the behavior of
Eastern Bluebirds feeding young, describe the diet of nestling and fledgling
bluebirds, and discuss the relationship between the foraging tactics of bluebirds
( Goldman 1975, Pinkowski 1977 j and types of prey fed to the young.

METHODS

Observations were made of Eastern Bluebirds nesting in nest boxes in Macomb Co.,
Michigan from 1971 to 1973. Nest sites were located in old fields adjacent to oak
iQuercus sp.) woodlands. Details of the study area are published elsewhere (Pinkowski
1975, 1976a). Relevant aspects of bluebird foraging were dealt with in a companion
paper (Pinkowski 1977).

I sampled 2503 nestling foods at 45 nests in 20 different nest sites and 275 fledgling
foods for 12 different broods. Animal foods were grouped into 23 taxonomic categories
(often families, occasionally orders or genera). I note individual prey species if these
appeared important and follow Cantrall (1968) and Kaston (1948) in assigning names
of various Orthopterans and spiders, respectively.

Nestling diet was sampled in part by using throat collars made from pipe-cleaners or
hea\y' thread to prevent the young from swallowing food. The collars did not appear to
harm the nestlings. Nests being sampled were checked every 20 to 30 min and young
were not deprived of food for more than 1.5 to 2.5 h per day. Rarely was the same nest
sampled on 2 consecutive days. Throat collars were difficult to use on small, recently-
hatched young unless an assistant held the bird while a collar was being applied.

The use of throat collars may generate results biased in favor of large items because
smaller items are likely to slip past the neck band ( Orians 1966 ) . To offset this bias
and enlarge the sample, I used 2 other methods of sampling foods: observations with a
spotting scope (15-60X1 and salvaging specimens (or portions thereof) from the nest
cavity or from the crops of dead nestlings. Salvaged specimens included food dropped by
the adults on trips to the nest and yielded small food items not likely to be obtained by
other methods. I found observing nests with a spotting scope useful on older nestlings
that could not be disturbed because of the possibility of premature fledging. This tech-
nicpie also permitted me to obtain a sample of 1359 foods fed by adults of known
sex (bluebirds are sexually dichromatic), and it was the only procedure used to sample
the food of fledglings. My presence 10-20 m from the nest did not disturb adult birds
feeding nestlings cr fledglings. Altogether, 54.3% of the nestling food data was obtained
by using a spotting scope, 36.9% by using throat collars, and 8.8% by salvaging specimens.

I sampled foods evenly throughout the day and nestling period to make the data
as representative of the diet as possible. Observations were conducted randomly to limit
interactions among variables. For example, nestlings of a given age were observed at

Pinkowski • FEEDING OF YOUNG BLUEBIRDS

different times of day to minimize the effects of diurnal variations in diet and feeding
rate.

Food items fed to the young are summarized as the percent occurrence of the various
taxonomic groups. Diurnal variation in prey and feeding rate was studied by assigning ac-
tivities to 1 of 4 time periods: early morning (06:00-10:00 EST j , late morning (10:00-
13:00), afternoon (13:00-16:00), and early evening (16:00-20:00) ; 28.9%, 32.3%, 20.7%,
and 18.1% of the nestling diet was sampled during the 4 time periods, respectively. Food
sampling activity was proportionate to the number of active nests and extended from 9
May to 15 August. Young of most first ( spring ) broods hatch in mid-May and fledge in
early June: second (summer) broods fledge between mid- July and mid- August (Pinkow-
ski 1976b). Nestling food samples were obtained on a monthly basis as follows: May,
41.5%; June, 22.5%; July, 25.8%; and August, 10.3%. Sampling was done under all
types of w'eather conditions, but results are slightly biased in favor of insects fed during
non-rainy conditions.

Vegetable matter is sporadic in the diet of nestlings and plant specimens found in the
nest cavity w^ere not necessarily fed to the young because the brooding parent may re-
gurgitate fruit seeds and skins (pers. obs.). For these reasons I analyzed the plant and
animal portions of the nestling diet separately. I included fruit in tabulations of the
fledgling diet because the limitations do not apply to young out of the nest.

Distances that adults foraged from the nest were recorded at 2 nests containing 3 and
5 young during the final week of the nestling period. Markers were placed in several direc-
tions at known intervals from the nest. Foraging bluebirds travel great distances and at
least 2 (often 3) observers communicating by radio were required to follow' the birds and
determine distances and directions at which prey w'as obtained relative to the nest.
Directions were placed in 1 of 16 categories (N, NNW, NW, etc.) for analysis of direc-
tional overlap by foraging adults.

Feeding rates are expressed in feedings per young per 15 h (=1 day) and represent the
average of results obtained for individual observation periods lasting 1-2 h (x = 86.5
min). I considered 1 trip to the nest with food as a single feeding regardless of the
number or size of the prey. The male bluebird, like males of some other passerines, may
offer food to the brooding female wdio in turn delivers it to the young. At some nests
70-90% of the nestlings’ food on the day of hatching is fed to them in this way. I con-
sidered food transfers, which become less common during the first w'eek and are rare
thereafter, as male feedings although the food is actually fed to the young by the female.

Frequency data, including the number of feedings of the male relative to the female,
were examined for significant differences by Chi-square. Differences in absolute feeding
rates (feedings young Ylay) were tested by a one-w'ay analysis of variance and Duncan’s
multiple range test (Steel and Torrie 1960:107). Diversity indices for prey taxa (H =
-2ii Piln Pi, where pi is the proportion of prey in the i^* taxon) were calculated from
information theory (Shannon and Weaver 1949). Because the diversity index is sensitive
to sample size ( Orians 1966, Pielou 1966) which in turn affects the number of prey
categories, I use this index only to compare groups having similar sample sizes.

RESULTS AND DISCUSSION

Rate of feeding nestlings. — Female bluebirds offered proportionately more
feedings to nestlings (54.8%) than males (45.2%). The difference is signifi-
cant (x“ = 19.0, P < 0.01, N = 2063 feedings), but considerable variation
existed from one nest to another.

THK WILSON BULI.ETIN • Vol. 90, Nu. 1, March 1978

l\()

70-

30 —

0 2 4 6 8 10 12 14 16 18 20

AGE IN DAYS

Fig. 1. Percentage of food contributed by male bluebirds and variation in feeding
fre<iuency of males and females combined during the nestling period. Data are based
on 168 h of observations f minimum: 5 h/nestling age). The line shows the significant
linear regression (P < 0.05) that existed for the first 17 days.

Feeding rate of both males and females did not depend on brood size. Males
averaged 6.4, 4.8, and 5.5 feedings h to nests containing 3, 4, and 5 young, re-
spectively. Corresponding figures for females are 6.5, 6.5, and 6.4 feedings h.
Conset}uently, young in nests containing 5 young received fewer feedings/day
( 35.5 ) than those in nests containing 4 young (42.2 I or 3 young ( 64.3) . That
feeding rate did not increase with brood size may in part reflect a reduction in
heat loss because of more insulation and less surface exposure in larger broods
( Mertens 1969 ) .

Bluebirds increased the feeding rate with nestling age during the first 17
days of the nestling period ( Fig. 1 I . During the first few days after hatching
there was an increase in prey size, and late in the nestling period adults oc-
casionally brought more than one item per trip to the nest. These changes
tended to offset the increase in feeding rate with nestling age.

3 he male and female contributed nearly erpial proportions of the nestlings’
food during the first 5 days of the nestling period (Fig. 1). Thereafter, the

Pinkowski • FEEDING OF YOUNG BLUEBIRDS

Diurnal Variation in

Table 1

Eeeding Rate of Eastern Bluebirds
1971-1973

IN Southeastern Michigan,

Male

Xo. Feedings

Female

Early Morning

224 (37.6%)

371 (62.4%)

Late Morning

281 (46.2%)

327 (53.8%)

Afternoon

202 (47.1%)

227 (52.9%)

Early Evening

225 (52.2%)

206 (47.8%)

female ceased brooding during the day (Pinkowski 1975 I and continued to
increase her feeding rate until day 16; after day 16 the female feeding rate
remained relatively constant (35.2-41.8 feedings young/day; x = 38.5). The
male feeding rate ( feedings/ young day I was low on days 1-5 ( x = 13.0 ) , was
significantly greater ( x = 24.8, P < 0.001 ) and fairly constant (range = 17.6-
28.8 j on days 6—17, and significantly decreased ( x = 14.9; P < 0.01) on days
18-21. Thus the overall increase in the feeding rate was at first attributable
to an increase in the feeding rate of both adults and later was due to an in-
creased rate by the female only. The decrease in the feeding rate late in the
nestling period was largely attributable to a reduction in feeding by the male.

Some male bluebirds ceased feeding the young soon after fledging. On 3
occasions males began new nests with different mates before young of the
previous nest were independent, a behavior not observed among females. In
such instances the female continued to feed the brood and supplied all of its
nutritional requirements.

Feeding rate (feedings/young day ) of males and females combined was
greatest in early morning (49.2), lowest in the afternoon (39.9), and nearly
identical in late morning and early evening (45.3 and 45.8, respectively).
Although none of the differences in feeding rates for the 4 time periods is
significant [P > 0.5 j, proportionately more feedings observed in the early
morning period (Table 1) were made by the female (^“ — 35.8, P < 0.001).
Also, males fed more and females fed less during the successive time periods;
the trend was significant (Z = 4.6, P < 0.001; Snedecor and Cochran 1967:
246 j . Thus there was a division of the daily “work load” by males and females
that may function to keep the number of feedings to the young relatively con-
stant throughout the day.

FOOD FED TO NESTLLXGS

Summary of invertebrate prey. — Lepidopterous larvae comprised the largest
percentage ( 32.4%) of animal food noted in the nestling diet and consisted of

THK WILSON BULLETIN • Vol. 90, No. 1, March 1978

HI]

sev(M'al families, includiiip: Noctuidae (“cutworms”), Arctiidae, Pieridae,
(ieomelridae, Nolodoiilidae, Pyralidae, and Si)hingidae. Adult Lepidoptera
accounted for T.6% of all animal foods recorded and consisted entirely of
moths I 1 lelerocera) .

Ortliopterans were the second largest group represented (25.6%), and in-
cluded grasshoppers ( Acrididae and one Tetrigidaej, 12.8%; crickets ( Gryl-
lidae, mostly the spring field cricket, Gryllus veletis; Alexander and Bigelow
1660 ), 9..‘1%; shield-hearing katydids ( Tettigoniidae: Decticinae; Atlanticus
testaceus) , 1.6%; various other katydids I Tettigoniidae exclusive of Dec-
licinae ) such as Neoconocephalus sp., Amhlycorypha sp., and Pterophylla sp.,
1.5%; and mantids (Mantidae, all nymphs), 0.4%. Spiders (Arachnida:
Araneae, including egg sacs, and a few Phalangida) were the third largest
group 1 11.3%), and generally consisted of wandering, ground-dwelling species
such as Lycosa frondicola.

Other taxa less freciuent in the nestling diet were beetles (Coleoptera; mostly
Phyllophaga sp., Melanotus sp., Scarites sp., and Cicindela sp. adults, and
Carabidae and Elateridae larvae), 11.0%; earthworms (Annelida: Oligo-

chaeta; Lumhricus sp. ) , 5.2%; various Hymenoptera (mostly carpenter ants,
Camponotus sp., and some Ichneumonidae) , 3.9%; and millipedes (Diplo-
poda) , 2.3%.

Food items uncommon in the nestling diet were: leafhoppers (Homoptera:
Cercopidae and Cicadellidae ) , 1.5%; sowbugs (Isopoda), 0.8%; snails and
snail shells (Pulmonata), 1.2%; flies ( Diptera) , 0.5% ; scorpion-flies (Mecop-
tera), 0.3%; dragonflies fOdonata: Anisoptera), 0.1%; Cicada. 0.1%; large
hugs (Hemiptera), 0.1%; and lacewings (Neuroptera: Chrysopida), 0.04%.
Unusual prey were 2 centipedes (Chilopoda, 0.1%) and 1 fairy shrimp
(Anostraca, 0.04% ).

Variations attributable to nestling age. — Spiders and Lepidoptera larvae
were the primary food of recent hatchlings (Table 2). As the young mature
more Orthoptera ( Gryllidae and Acrididae), Coleoptera, and earthworms were
fed. Prey diversity was lower early in the nestling period (H = 1.60 for young
1-5 days old) than later ( H = 2.05 and 2.04 for young 6-10 and 11—18 days
old, resi)ectively ) .

Nine of 12 food items fed to young 1 day old or less were spiders. Twelve
spider species were noted only once during the sampling period; 9 of these
species occurred only in the diet of nestlings 4 days old or less. Other pas-
serines also exhibit a preference to feed spiders to recent hatchlings I Royama
1970). Small nestlings must he fed small, easily digested foods, and prey with
a high energy content relative to its size would seem most desirable. Spiders
have a soft abdomen, lack coarse appendages, and have greater caloric

Pinkowski • FEEDING OF YOUNG BLUEBIRDS

Frequency of Animal

Foods

Table 2

Fed to Nestling Eastern Bluebirds
Michigan, 1971-1973

IN Southeastern

0-5 Days Old

6-10 Days Old

11-18 Days Old

No.

Lepidoptera larvae

163

41.6

192

35.0

402

39.1

Arachnida

121

30.9

11.5

6.8

Acrididae

6.9

15.7

171

16.6

Gryllidae

7.4

10.2

8.4

Coleoptera

1.0

5.1

4.7

Heterocera adults

6.6

6.0

3.0

Hymenoptera

1.0

7.3

4.9

Lumbricus sp.

1.0

2.2

7.9

Atlanticus testaceus

0.8

2.4

2.0

Tettigoniidae

0.5

2.4

equivalents than Acridids and earthworms (cal/g dry wt; Golley 1961, Van
Hook 1971 ) that bluebirds feed more often to older nestlings.

Large spiders (e.g., Lycosa frondicola and Schizocosa avida) were noted in
the diet of older nestlings, and male and female spiders of dimorphic species
were selected by size for young of different ages. Eleven L. frondicola males
were fed to nestlings averaging 5.8 days old, whereas 7 females of this species
(which are larger than males) were fed to young an average of 7.1 days old.
A similar trend appears among some Orthoptera; nymphs of the grasshopper
Melaiioplus bivittatus were fed to 3-day-old nestlings and the coarser adults
were not fed until day 7 (males, which are smaller than females) and day 9
(females) .

Variations attributable to season. — Invertebrate prey fed to nestlings and
tabulated on a monthly basis revealed that spiders were fed more in May and
June (13.1% and 14.3%, respectively) than July (7.9%) and August (4.4%).
Ihe seasonal decline in frequency of spiders was not entirely attributable to a
decrease in availability. L. frondicola and Phidippus princeps, the most com-
mon spiders in the nestling diet in spring, were not fed after early June (Fig.
2) although both species are present from April to October at the latitude of
my study area (Dondale 1971).

Ground-dwelling spiders belonging to the family Lycosidae (e.g., L.
frondicola.^ Trochosa terricola) were more common in the diet of nestlings
in spring. In summer, however, spiders of the family Thomisidae (e.g.,
Tibelliis oblongus, Xysticus elegans) that dwell on herbaceous plants and tree
trunks (Lowrie 1948) were more common. The Lycosid Schizocosa avida

THE WILSON BULLETIN • Vol. 90, No. 1, March 1978

' MAY ' JUNE * JULY ' AUG. '

Fig. 2. Seasonal variation of spider taxa in the diet of nestling bluebirds. Numerals
indicate sample sizes. All taxa noted at least 3 times are included. The solid portion of
the lime scale indicates the sampling period. The species are represented as follows:
Trochosa terricohi (Tt), Thanatus jormicinus (Tf), Lycosa jrondicola ILf), Phidippus
princeps ( Pp ) , elegans (Xe) , Schizocosa avidu (Sa),and Tibellus oblongus (To).

is common in summer but this species, like the Thomisids and unlike the other
Lycosids, is phytophilous ( Kuenzler 1958).

Lepidoptera adults (all moths) were more common in the nestling diet
in summer (5.6% ) than in spring (2.9%). Earthworms and Coleoptera were
staple food items in spring, especially during rainy periods, but became less
important later in the season. Earthworms comprised 10.8% of the nestling
diet in May and 3.3% in June, but were absent after 1 July. Coleoptera com-
prised 17.9% of the diet in May, 8.3% in June, 5.9% in July, and 0.8% in
August. Hymenoptera were more common in May (4.7%) and June (4.8%),
when swarming carpenter ants were frequently taken, and less common in
July ( 2.6% j and August (2.4%).

Lepidoptera larvae were more common in the diet during May (35.4%) and
June (41.8%) than July (20.8%) and August (28.9%). As was the case
for spiders, seasonal changes in occurrence of larval Lepidoptera reflected
changes in availability, but geophilous forms were more common early in the
season. Cutworms (Noctuidae larvae) accounted for 46.6% (N = 393) of
all Lepidoptera larvae noted in the diet. One species, the bronzed cutworm
{ IVephelodes minians), comprised 48.6% of the cutworms recorded and is
typical of the prey belonging to this taxon in that it feeds at night but is
found on the ground during the day. The percentage of cutworms among all
Lepidoptera larvae fed to nestlings was greatest in May (74.6%, N = 134) and
decreased in June (36.4%, N = 140), July (33.7%, N = 83), and August

Pinkowski • FEEDING OF YOUNG BLUEBIRDS

- Mv

- Cv
As

227

-N

“ Ms
Pb

* MAY JUNE JULY AUG. '

Fig. 3. Seasonal variation of Orthoptera taxa in the diet of nestling bluebirds. Numer-
als indicate sample sizes. All taxa noted at least 3 times are included. The solid portion
of the time scale indicates the sampling period. Individual species and genera are repre-
sented as follows: Gryllus sp. iG) , Pardalophora apiculata iFa) , Arphia sulphurea (As),
Chortophaga viridifasciata (Cv), Melanoplus viridipes (Mv), Atlanticus testaceus (At),
Pseudopomala brachyptera (Pb), Melanoplus sanguinipes (Ms). Melanoplus hivittatus
(Mb), Melanoplus confiisus (Me), Chorthippus curtipennis (Cc), Dissosteira Carolina
i\)c) , Neoconocephalus sp. (N ), and Mantis sp. (M).

(11.4%, N = 35). Larvae of other Lepidoptera families ( e.g., Geometridae)
that inhabit trees and vegetation became increasingly common as the season
progressed, but maximum consumption of all families combined occurred in
June.

Orthoptera were more common in the diet in summer than spring, although
individual species recorded were dependent on season (Fig. 3). Gryllus
veletis, the most common Orthoptera noted, increased steadily from May
through July (5.1%, 8.4%, and 17.6% for the 3 months, respectively). G.
veletis nymphs were fed in mid-May, adults in late May, and peak predation
occurred in mid- July. Acrididae increased steadily from ^lay to August (5.5%,
9.2%, 23.4%, and 26.9% for each month, respectively ) .

TllK WILSON lUILLKTIN • VoL 90, No. I, March 1978

The incidence of various Orthoplera in the nestling diet differed little from
the chronological appearance of the various species in the study area (pers.
ohs. ; Canlrall 196o ). Orlhopteran species of similar size and habits, however,
complemented each other in the diet on a seasonal basis. In summer adults of
Atlanticus testaceus, a large, geophilous species, replaced G. veletis in the diet.
Large vernal Acridids that overwinter as nymphs in southern Michigan
iChortophaga viridijasciata and Pardalophora apiculata) were replaced in
summer by other large Acridids ( Dissosteira Carolina, Melanoplus hivittatus).
Similarly, the smaller Acridids Arphia sulphur ea and Melanoplus viridipes
were common in si)ring and were replaced in summer by Pseudoponiala hra-
chrptera, Chorthippus curtipennis, Melanoplus confusus, and M. sanguinipes,
which are also small.

Phytophilous Orthoptera [Neoconocephalus sp. and Mantis sp.) were fed
to nestlings only in summer. Thus for all 3 of the major prey groups
( Lei)idoptera larvae. Arachnids, and Orthoptera j, bluebirds tended to select
geophilous species in spring and phytophilous species in summer. Phytophil-
ous invertebrates were undoubtedly more abundant relative to geophilous taxa
late in the season as vegetation height increased, but in some cases geophilous
prey were present late in the season, but were ignored by bluebirds.

Evans (1964 ) found that Vesper Sparrows (Pooecetes gramineus) , Song
Sparrows [Melospiza melodia), and Chipping Sparrows [Spizella passerina)
breeding in southern Michigan use a greater variety of food in summer than
spring. I found this somewhat true of bluebirds; 19 of the 23 (82.6%) prey
categories were represented during July whereas only 14 (60.9%) were re-
corded in May. Diversity indices were higher in July (2.24) and August
(2.09) than in May ( 1.95 I and June (1.96).

Variations attributable to time of day. — Several classes of prey, including
Arachnida, Coleoptera, Heterocera, and Tettigoniidae, displayed no frequency
variation with time of day; others, however, were more variable. Gryllidae
were fed more in early morning (13.7%) and early evening ( 13.5%) than late
morning (4.9%) and afternoon (4.2%). Acrididae displayed the reverse
pattern ( 19.0% in the afternoon, 18.7% in the late morning, 12.8% in early
evening, and 8.0% in early morning ) . Thus both Gryllidae and Acrididae
were apparently preyed upon most often when they were most active.

Lepidoptera larvae were abundant (39.2-44.7%) from early morning until
late afternoon and less common (29.2%) in early evening. Hymenoptera
were most abundant in early evening ( 8.3%) when bluebirds frequently engage
in flycatching ( l^inkowski 1977); they were least common in the afternoon
(0.6%) and intermediate (4-5%) in the other periods. Earthworms were
most common in early evening (8.7%) and afternoon (7.3%), and less com-
mon ( 1. 6-3.0%) in other periods.

Pinkowski • FEEDING OF YOUNG BLUEBIRDS

Prey diversity was greatest in early evening (H = 2.18), partly because
aerial insects ( Hymenoptera, Diptera) were added to the diet at that time.
Diversity was lowest in the afternoon (H = 1.72 ) when feeding rate was re-
duced, and was greater in late morning (H = 1.93) and early morning (H =
1.90).

Small stones, snails, and snail shells function as grit (Royama 1970) and
were noted only in the early morning. The female bluebird apparently sup-
plies nearly all of the grit required by nestlings. Seven observed feedings of
grit were all made by the female.

Variations attributable to weather. — Precipitation (mostly rainfall except
during March ) increased during the 3 years of study; 15.3 cm of precipitation
fell from 1 March to 30 June 1971, and 25.6 cm and 38.6 cm were recorded for
the same period in 1972 and 1973, respectively. Annual incidence of Acrididae
in the diet decreased with the increasing precipitation (20.1%, 16.1%, and
4.7% for the 3 consecutive years) ; the same trend occurred among Gryllidae
( 17.1%, 8.5%, and 4.9%). Some Orthoptera, especially grasshoppers, flourish
during periods of drought and are reduced in numbers during rainy years
(Shelford 1963:318, Scharff 1954).

More earthworms were taken in 1973 (14.0%) than in 1971 (2.9%) and
1972 (1.5%). A paucity of Lepidoptera larvae in the diet in 1971 (19.9%)
compared with 1973 (35.7%) and 1972 (38.2%) may have been attributable
to death of these insects from desiccation during dry conditions ( Andrewartha
and Birch 1960 ) or other factors such as lack of food. In any event, bluebirds
tend to feed Orthoptera during dry seasons and Lepidoptera larvae and
earthworms during rainy seasons, presumably because of differences in rela-
tive availability.

Fruit fed to nestlings. — Vegetable matter, uncommon in the diet of nestlings,
was noted at only 4 of 45 nests observed. The fruits involved were mulberries
{Morus sp.j, raspberries (Rubus sp.j, dogwood iCornus stolonifera) , cherry
{Prunus virginiana) , and honeysuckle {Lonieera sp. ). Fruit was not fed be-
fore late June, when it became abundant in the study area. At 2 of the 4 nests,
each containing nestlings within a few days of fledging, fruit comprised 33.0%
and 37.0% of the nestling diet over 3 and 5 day periods, respectively ( approxi-
mately 15 h observation in each case ) . At the 2 other nests fruit was noted
only once; each instance involved older nestlings (^14 days old).

Morton ( 1973 j concluded that a fruit diet prolongs nestling development
and is selected against as a food for poikilothermic young on account of its
low protein content. The altricial strategy, he argues, re(iuires that the small
young be able to use food principally for growth and not for heat production
(because the nestlings’ heat requirements are satisfied by lirooding). Fruit,
therefore, is not a dietary constituent of young l)luel)irds until the last week of

THE WILSON BULLETIN • VoL 90, No. 1, March 1978

the nesllinp period, when the nestlings are completely endothermic (Pinkowski
1975).

Partitioning oj the feeding niche. — A foraging pair of adult birds may re-
duce competition hy differentially using the feeding resources available to
them without necessarily involving secondary sexual dimorphism ( Ligon
1968, Jackson 19701. I noted no difference in the diversity of foods fed to
the young hy male (H = 1.89) and female (H = 1.87 j bluebirds. Males,
bowever, fed significantly more Gryllidae (x“ — 4.9, P < 0.05) and earth-
worms (x“ = 21.9, P < 0.001 ) than females; females fed more Acrididae (x“
= 5.4, P < 0.05), Hymenoptera (;(“ = 4.2, P < 0.05), and Arachnida (x“ =
4.9, P < 0.05) than males. Little difference was noted among other prey
categories including moths = 0.8, P > 0.3) and Lepidoptera larvae iP
> 0.9).

I could not attribute differences in foods fed by males and females to dif-
ferent feeding rates of males and females relative to age of the nestlings.
Earthworms and crickets, preferred foods of males, were most common in the
diet of older nestlings that were fed more by females. Spiders were relatively
uncommon late in the nestling period when females fed more often than males.
Although grasshoppers were common in the diet of older nestlings, Pinkowski
(1974 ) noted that captive female Eastern Bluebirds and Mountain Bluebirds
{Sialia currucoides) preferred to feed grasshoppers to nestlings.

Differential prey use may result from differential use of the feeding
range by males and females. Using pooled data for 2 nests, I found that
male bluebirds obtained prey for nestlings closer to the nest site ( x = 113.6 m,
SD = 99.4, N = 256) than females ( x = 152.4 m, SD = 117.3, N = 182; t
= 3.7, P < 0.01 j. Indices of overlap (Horn 1966 I for directions that males
and females obtained prey were great (0.875 and 0.902); apparently food
resources were not partitioned on a directional basis.

In some areas male and female bluebirds forage at equal distances from the
nest ( Pinkowski 1974, Goldman 1975). When there is a difference in foraging
distances, however, evidently the male remains closer to the nest, possibly
because male bluebirds play a greater role than females in defence of the nest
cavity against conspecific intruders. Females of some open-nesting species
such as the Bobolink { Dolichonyx oryzivorus) and Henslow’s Sparrow {Am-
niodramus hensloivii) forage closer to the nest than males (Wiens 1969, Robins
1971).

Power ( 1974:88-99 ) related foraging distance of adults to brood size (work
load ) in the Mountain Bluebird. For the 2 Eastern Bluebird nests 1 examined,
however, the adults with 3 young foraged farther from the nest ( x = 166.2 m )
than adults with 5 young ( x = 96.2 m; t = 7.1, P < 0.001). Eastern Bluebirds
are more (lei)endent on feeding perches than Mountain Bluebirds and are

Pinkowski • FEEDING OF YOUNG BLUEBIRDS

known to vary foraging range according to perch distribution ( Pinkowski
1974, 1977). Thus habitat quality appears more important than the number of
young in the nest in determining how far adults travel in search of food.

FEEDING OF FLEDGLINGS

Foods fed to fledglings differ from those fed to nestlings. Lepidopterous
larvae were more common in the fledgling diet (44.0% of all fledgling foods
recorded), reflecting peak consumption in June when most fledglings were
out of the nest. Earthworms (11.4%) and Coleoptera (7.7%) were more
common in the fledgling diet than the nestling diet, hut the reverse was true for
Acrididae (8.8%), Arachnida (4.4%), Gryllidae (3.3%), and Heterocera
(3.3%). Fruit (mulberries and cherries) comprised 11.0% of the fledgling
diet, but was noted only during the summer period ( July and August) . General
observations indicated that the adults feed smaller items to fledglings than
nestlings.

Feeding patterns of adults foraging for fledglings differ from those of adults
feeding nestlings. Fledgling bluebirds spend most of their time in large trees
and alternate active and inactive periods; they begin calling when hungry and,
depending on food availability, receive several feedings until satiated. Adults
obtain many food items within a few meters of the fledglings, often by gleaning
from the tree tops, and many small items may be fed in rapid succession to
young out of the nest. This is in contrast to the long trips with large items
made regularly by adults with young in the nest.

CONCLUSIONS

Prey availability is important in determining dietary constituents of young
bluebirds. Weather and time of day influence prey activity and abundance
and hence affect what is fed to the young. The presence of smaller nestlings
somewhat restricts prey selection because older young are fed a greater variety
of foods. As the spectrum of suitable prey increases with nestling age, how-
ever, so does the amount of food required by the young and consequently the
feeding rate of adults. These changes would tend to equalize the time and
energy expended by adults during the duration of the nestling period.

The data obtained in this study corroborate Goldman’s (1975) conclusion
that bluebirds feed large food items to nestlings. Lepidoptera larvae (especially
cutworms) are the preferred food for nestlings. Beal (1915), however, states
that Orthoptera are preferred by adult bluebirds and noted that Coleoptera are
nearly twice as common in the diet of adults (29.9%) as I found in the diet of
nestlings. Although Orthoptera and Coleoptera are large, their relative in-
frequency in the diet of the young may he explained by their coarseness.

THE WILSON BULLETIN • Vul. 90, No. 1, March 1978

Coarse foods re(iuire more preparation and thereby reduce caloric yield per
unit time, the basic determinant of food value ( Emlen 1906 j.

Forafi:inf>: bluebirds locate prey from a distance by using conspicuous feeding
perches; in sining most prey is obtained after a short “drop” to the ground, but
in summer there is an increased use of tactics such as gleaning and flycatching
that result in i)rey capture above ground ( Pinkowski 1977). Analysis of
seasonal variation in i)iey taxa suggests that the seasonal trend in foraging
tactics is independently related to both a seasonal increase in vegetation bio-
mass ( height and density ) and an increase in the abundance of invertebrates
living above ground. Bluebirds rarely feed by dropping onto the ground in
areas having tall, dense vegetation, probably because doing so would often re-
quire them to relocate prey from close range and not from a conspicuous and
elevated position (Pinkowski 1974:66). Thus late in the season bluebirds do
not feed upon some geophilous prey taxa (earthworms, cutworms, Coleoptera,
and some spiders ) that are still available, but instead exploit phytophilous and
aerial prey (moths and certain spiders, Lepidoptera larvae, and Orthoptera )
that are more abundant and conspicuous from a distance than geophilous prey.
By changing their predatory tactics on a seasonal basis, bluebirds are able to
exploit changes in prey availability as well as maintain the optimum predatory
efficiency permitted by their perch-feeding habit.

SUMMARY

The behavior of adult Eastern Rluel)irds feeding nestlings and fledglings and the diet
of young bluebirds were studied in southeastern Michigan from 1971 to 1973. Females
fed nestlings more often than males. The feeding frecjnency increased with nestling age
until just prior to fledging, when a decline occurred. Feeding rate of males and females
combined was relatively constant throughout the day although females fed young more
often earlier in the day and male feeding rate was greater later in the day.

Lepidoptera larvae were the most common food of both nestlings and fledglings and
comprised 32.4% of the nestling diet. Orthoptera (mostly Acrididae and Gryllidae) were
also common (25.6%), especially in summer. Spiders (11.3%) were particularly im-
portant early in the season and for newly-hatched young. Fruit was uncommon in the
diet of nestlings hut was fed to fledglings in summer and made up 11.0% of all fledgling
foods recorded.

Adult males and females fed different foods to the young, thereby partitioning the feed-
ing niche. Males fed significantly larger percentages of Gryllidae and earthworms;
females fed larger percentages of Arachnida and Acrididae.

ACKNOWLEDGMENTS

Portions of this paper originally comprised part of a Ph.l). dissertation submitted to
the Department of Biology, Wayne State University. William Thompson, Claude Rogers.
Melvin W eisbart, .‘Stanley (iangwere, and Diane Pick read the manuscript and made help-
ful suggestions and criticisms. (Charles Dondale identified the spiders, and Michael
Tyrkus, John Newman, and Eric Ouinter helped identify many of the insects. James

Pinkowski • FEEDING OF YOUNG BLUEBIRDS

Stevens, Patrick Pinkowski, and my wife, Pliyllis, gave unselfish assistance in the field.

I am grateful for the help of all these persons.

LITERATURE CITED

Alexander, R. D. and R. S. Bigelow. 1960. Allochronic speciation in field crickets, and
a new species, Acheta veletis. Evolution 14:334-346.

Andrew AHTHA, H. G. and L. C. Birch. 1960. Some recent contributions to the study of
the distribution and abundance of insects. Annu. Rev. Entomol. 5:219-242.

Beal, F. E. L. 1915. Food habits of the thrushes of the United States. U.S.D.A. Biol.
Surv. Bull. 280.

Cantrall, I. J. 1968. An annotated list of the Dermaptera, Dictyoptera, Phasmatoptera,
and Orthoptera of Michigan. Mich. Entomol. 1:299-346.

CoTTAM, C. AND P. Knappen. 1939. Food of some uncommon North American birds.
Auk 56:138-169.

Davison, V. E. 1962. What bluebirds eat. And. Mag. 64:223-224.

Dondale, C. D. 1971. Spiders of Heasman’s field, a mown meadow near Belleville,
Ontario. Proc. Entomol. Soc. Ontario 101 :62-69.

Emlen, J. M. 1966. The role of time and energy in food preference. Am, Nat. 100:611-
617.

Evans, F. C. 1964. The food of Vesper, Field, and Chipping sparrows nesting in an
abandoned field in southeastern Michigan. Am. Midi. Nat. 72:57-75.

Forbes, S. A. 1903. On the food of birds. Bull. 111. Lab. Nat. Hist. 1:86-162.

Goldman, P. 1975. Hunting behavior of Eastern Bluebirds. Auk 92:798-801.

Golley, F. B. 1961. Energy values of ecological materials. Ecology 42:581-584.

Horn, H. S. 1966. Measurement of “overlap” in comparative ecological studies. Am.
Nat. 100:419-424.

Jackson, J. A. 1970. A quantitative study of the foraging ecology of Downy Wood-
peckers. Ecology 51 :318-323.

Kaston, B. j, 1948. Spiders of Connecticut. Conn. State Geol. Nat. Hist. Surv. Bull.
70:1-874.

Kuenzler, E. L. 1958. Niche relations of three species of lycosid spiders. Ecology 39:
494-500.

Eicon, J. D. 1968. Sexual differences in foraging behavior in two species of Den-
drocopos woodpeckers. Auk 85:203-215.

Lowrie, D. C. 1948. The ecological succession of spiders of the Chicago area dunes.
Ecology 29:334-351.

Mertens, j. a. L. 1969. The influence of brood size on the energy metabolism and
water loss of nestling Great Tits, Pams major major. Ibis 111:11-16.

Morton, E. S. 1973. On the evolutionary advantages and disadvantages of fruit eating
in tropical birds. Am. Nat. 107:8-22.

Orians, G. H. 1966. Food of nestling Yellow-headed Blackbirds, Cariboo Parklands,
British Columbia. Condor 68:321-337.

PiELOu, E. C. 1966. The measurement of diversity in different types of biological col-
lections. J. Theoret. Biol. 13:131-144.

Pinkowski, B. C. 1974. A comparative study of the liehavioral and breeding ecology of
the Eastern Bluebird iSialia sialis). Ph.D. thesis, Wayne State Univ., Detroit, Mich.

— 1975. Growth and development of Eastern Bluebirds. Bird-Banding 46:273-289,

-. 1976a. Use of tree cavities by nesting Eastern Bluebirds. J. V ildl. Manage.

40:556-563.

THK WILSON BULLETIN • Vul. 90, No. 1, March 1978

- — 19761). IMiotoperiodic (‘ffects on tlie postjuvenal molt of the Eastern Bluebird.

Ohio .). Sci. 76:268-273.

. 1977. Foraging l>ehavior of the Eastern Bluebird. Wilson Bull. 89: 404-414.

PowKR, H, W . 1974. The Mountain Bluebird: sex and the evolution of foraging be-

havior. Ph.l). thesis, Univ. .Mich., .Ann .Arbor, Mich.

Robins, J. 1). 1971. Differential niche utilization in a grassland sparrow. Ecology 52:

1065-1070.

Royama, T. 1970. Factors governing the hunting behavior and selection of food by tbe
(ireat Tit {Pants major L. I . J. .Anim. Ecol. 39:619 668.

I*'CH \BFF, 1). K. 1954. Tbe role of food jilants and weather in the ecology of Melanoplus
rnexicanus mexicaniis (Sauss. ). J. Econ. Entomol. 47:485-489.

Shannon, C. E. and W. Weaver. 1949. The mathematical theory of communication.
L niv. of 111. Press, L rbana.

.Shei.ford, E. 1963. The ecology of North .America. Univ. of 111. Press, Urbana.

Snedecor, G. W. and W. G. Cochran. 1967. Statistical metbods. Iowa State Univ.
Press, Ames.

Steel, R. G. 1). and J. H. Torrie. 1960. Principles and procedures of statistics. Mc-
Graw-Hill, New York.

\’an Hook, R. L, Jr. 1971. Energy and nutrient dynamics of spider and orthopteran
populations in a grassland ecosystem. Ecol. Monogr. 41:1-26.

Vi lENS, J. .A. 1969. .An approach to the study of ecological relationships among grass-

land birds. Ornithol. Alonogr. 8:1-93.

245 COUNTY LINE RO.\D, BRIDGEVILLE, P.\ 15017. .ACCEPTED 11 NOV. 1976.

DIFFERENTIAL USE OE ERESH WATER ENVIRONMENTS BY
WINTERING WATEREOWL OF COASTAL TEXAS

Donald H. White and Douglas James

Species having similar life styles (Ralph 1975) characteristically occupy
different ecological niches (Hutchinson 1957, 1965) within shared environ-
ments. Many workers have shown that this principle seems to he operative in
avian communities ( MacArthur 1958, Cody 1968, James 1971, Posey 1974,
Whitmore 1975 ). Our study was conducted to determine how feeding flocks
of wintering waterfowl coexisted in feeding site selection, what environmental
factors that were measured were the most important in certain aspects of niche
separation, and how the niches were arranged in the aquatic community at
the study site.

METHODS AND MATERIALS

Study area. — Data were collected from early October through late December 1973 from
2 adjacent ox-how lakes on the grounds of the Welder Wildlife Foundation in San
Patricio County near Sinton, Texas. These fresh water lakes were up to 2.5 m deep hut
averaged about 1.5 m in the middle. A broad zone of semi-aquatic grasses ( Paspalum and
Panicam) occupied the perimeters and hurhead ( Echinodoriis rostratus), southern cut-
grass (Leersia hexandra) , and bulrush i Scirpus californicus) occurred in isolated small
patches. The transition zone from emergent semi-aquatic vegetation sometimes occurred
over 90 m from shore, hut was (juite variable in position. Extensive floating or partly sub-
merged patches of acjuatic vegetation were dominated by southern naiad iNajas
guadahipensis) , star grass ( Heteranthera liebmannii) , musk grass iChara), and duck
weed (Lernna perpiisilla) .

Large numbers of waterfowl use the coastal region of southern Texas during the fall
and winter months ( Bellrose 1976) therefore, references to “wintering waterfowl” and
“wintering grounds” throughout this paper are made on this basis. Most of the individuals
of some species such as the Fulvous Whistling Duck and Blue-winged Teal have
moved further south by late December or early January (Bennett 1938, Bellrose 1976) and
may not be considered as truly wintering species of southern Texas. Nevertheless, these
2 species were included as they were present when the study was conducted.

Field methods. — The species studied were; Mottled Duck (Anas fitlvigula), Pintail
(Anas acuta), Gadwall (Anas strepera), American Wigeon (Alias americana) , Northern
Shoveler (Anas clypeata). Blue-winged Teal (Anas discors), Green-winged Teal (Anas
crecca). Fulvous Whistling Duck ( Dendrocygna bicolor). Redhead (Aythya americana) ,
Canvashack (Aythya valisineria) , Ring-necked Duck (Aythya collaris) , Lesser Scaup
(Aythya ajjinis). Ruddy Duck (Oxyura jamaicensis) , and American Coot (Fulica ameri-
cana).

To characterize the environments of feeding waterfowl flocks 20 factors were measured
in the field comprising social, vegetational, physical, and chemical properties. Social
factors included: total number of ducks in flock, number of species in flock, number in
flock of species being sampled, number feeding of species being sampled, number of

100

THK W II.SON nULLETIN • Vul. 90, No. I, March 1978

coots present in flock, and distanct! to neighboring flock On). Vegetational factors were:
% (‘incrgent vegetation, enn'rgent vegetation height (cm), and % floating and or sub-
merged vegetation. Physical factors were: depth of water at feeding site (cm), distance
of flock from shore (m), turbidity of water at feeding site (Jackson turbidity units), %
cloud cover, and wind velocity (km hr). The chemical measurements of the water at
feeding locations were: pH, dissolved oxygen (pj)in), total nitrogen (ppm), total pbos-
phorous (ppm), total calcium (ppm), and conductivity (micrombos/cm ) .

J'wenty-five samples of the 20 environmental factors were measured for each species.
Feeding flocks were sampled at random and data collecting for each species was distributed
as much as possible during the study period to eliminate time of sampling as a bias. Also,
60 random samples of the environmental factors (excluding social factors) were taken
to determine the general nature of the habitat available in the aquatic environment at
Welder. The random samples were selected by superimposing a grid on a map of the
study area and using numbers from a random table as X and Y coordinates to designate
approximate sample locations. Means and standard deviations of the factors measured in
the study for each species and the random habitat samples are included in White (1975).

The feeding flocks of wintering waterfowl were studied regardless of size. Although
loose mixed-species flocks often were encountered, the ducks tended to separate according
to species. Therefore, the approximate center of each species flock within loose mixed
flocks served as the sample point from which measurements were made. Sampling began
at daylight and continued throughout the day. A canoe and hip boots were used in
collecting data. Observations were made with binoculars and a telescope. Social factors
were recorded from afar and the location of nearest neighboring flocks was noted before
disturbing the ducks to measure other factors.

Flock-center locations were marked using a buoy and samples were taken within a
radius of approximately 3 m from this point. Percentages of emergent vegetation and
floating and or submerged vegetation were estimated by making 50 random observations
within the sampling perimeter using a sighting tube (^ inkworth and Goodall 1962) and
doubling the total sightings having plants intersected by crosshairs.

Water depth was measured with a meterstick or weighted nylon cord; distances to shore
and to nearest neighboring flock were measured with a range finder; wind velocity
was measured with an anemometer held at eye level; cloud cover was estimated. A
water sample was taken at each site and analyzed at the end of the day for turbidity and
chemical factors using a Hach water analysis kit.

Population densities of the waterfowl species included in this study were highly variable.
For example. Pintails generally were much more abundant than Mottled Ducks, Canvas-
backs, or Ruddy Ducks. Total numbers of the various species using the lakes at W elder
varied from day to day since waterfowl are highly mobile and may cover a wdde range
of habitats. Certainly it is possible that on one or several occasions measurements were
taken on the same individuals of a particular species. This should not bias the data
(James 1971) since individuals of a species generally are indicative of that species as a
whole. Population estimates for the 3 month period are not available per se, however see
W hite ( 1975) for mean flock sizes l)ased on 25 observations for each species.

Data analysis. — The IBM-360 Model 50 digital computer at the University of Arkansas
was used for all data analyses. Principal component (PC) analysis (Morrison 1967) based
on correlations between untransformed data was used to determine the environmental
factors that varied the most in niche relationships. After transforming the data to
minimize heteroscedasticity and non-normality (Box and Cox 1964, Andrews et al. 1971).
multivariate analysis of variance (Cooley and I.ohnes 1971) with a step-down procedure

White ami James • WATERFOWL HABITAT USE

101

( Bargnian 1962) was used to determine how the species were arranged with respect to the
environmental factors that were important in separating species. The canonical scores
from the preceding analysis were subjected to 1-way analysis of variance with Duncan’s
multiple range test (Steel and Torrie 1960) to determine the degree of species environ-
mental overlap.

RESULTS

The following PC analyses were conducted on the combined species data.
The first included all of the 20 environmental factors measured in the study;
the second involved only the 14 non-social factors (vegetational, physical, and
chemical ) . In both analyses the initial principal components identified the
combination of factors that described the greatest variation in the data sets.
This represented the breadths and limits of the ecological niches based on the
factors that were measured. Niche differences were evaluated using multi-
variate analysis of variance and associated procedures.

Overall relationships. — The PC analysis that included all of the 20 environ-
mental factors measured in the study gave an overall account of niche structure
for the species, including the social environment as a niche component. The
first principal component (PC-I) of the combined data set for all species
showed high correlation values for 4 social factors (Table 1). This indicated
that waterfowl as a group varied the most in social activity. The second prin-
cipal component ( PC-II ) showed high correlation values for water depth at
feeding site, vegetational percentages, calcium content, and conductivity.
Combinations of these factors characterize specific feeding sites. Together
PC-I and PC-II accounted lor 30% of the total environmental variance.

A 2-dimensional representation of the distribution of the ecological niches
(Fig. Ij was produced by plotting the mean PC-I and PC-II scores (James
1971 j. Relative niche widths are shown by 1% confidence ellipses circum-
scribing the mean of each species data set. The ellipses are very small indi-
cators of niche width; larger ellipses would tend to mask relationships due to
broad overlap. Social activity, based on those social factors with high correla-
tion values in Table 1, increases from left to right along the PC-I axis (Fig. 1).
Water depth at feeding site and floating and or submerged vegetation increase
from top to bottom along the PC-II axis, whereas calcium and conductivity
(high values equated to high productivity; Orians 1966, Russell-Hunter 1970)
and emergent vegetation decrease in the same direction. Each species position
within the total environmental space is determined by its individual responses
to the definitive factors characterizing the space.

The Redhead and Canvasback were quite similar in response and exhibited
the most social activity, whereas the Mottled Duck was the least social ( Fig. T) .
The Ruddy Duck and Gadwall generally occupied the deeper water with
copious acjuatic vegetation (Sincock 1963, Bellrose 1976) while at the other

Table 1

COKKKLATIONS WlTII FlKST AND SeCOND PRINCIPAL COMPONENTS lUSEI) On AlL ENVIRONMENTAL FACTORS AND ON AlL ExCEPT THE SOCIAL

Environment

102

THE WII.SON lUJEEETIN • Vol. <J0, No. 1, March 1978

m i-H oc o^ ^

I— _ lO _ O O CO

CO CC CN t- O CO lO t'- IC o^ vC
VC CC <—l sC vO n- I— Csj LO LO O r-

LOOOCsCvOOa

eg cs} eg ej co o ej t- o

S 13. 4 ^ g ^ S § S

•^COCOCOOO lOcolO VC

i--C'coocor- oo'— ot^ocer'- vcccegcvioco
cq eg cq r-. to. lo eg co. eg r- o o eg r-n o o o eg

c«

1“ j

$ 1

■s

III

c _ JO ~

^ ^ O - • -

S B Z'^1 Z

0 ^ ^ J ^

X a 5 S E £

i: n E

”1=

= .2

C Z

Z it

C 4,

iH >

= 3

^ 2

2 fcc

Z o
£ t

W 2-

£ ^

^ ^ C/)

I ? ^

> , "Z

£ E
^ c =

5 iC C, 3

■glil

> c =1, o

= X .=

, - — C

— _= = .= ? _ -=

4- .X c :i^ •- c c o

I— ^ eg ^ LO vc"

o ^ eg CO ■r}>'

sc CO ce o

White and James • WATERFOWL HABITAT USE

103

Fig. 1. Waterfowl ordination with 1% confidence ellipses based on species values and
means for scores of the first (abscissa) and second (ordinate) principal components;
social, vegetational, physical, and chemical environmental factors included in the analysis.

extreme the Green-winged Teal and Mottled Duck favored shallow productive
waters with much emergent vegetation ( Singleton 1968, Bellrose 1976 } .
Grouped ellipses show similarities in mean niche characteristics of various
waterfowl such as the closeness in the Blue-winged Teal, Northern Shoveler,
and Ring-necked Duck. The American Coot and Fulvous Whistling Duck had
the largest niche sizes, the Ring-necked Duck and Ruddy Duck the smallest.
The Redhead and Canvasback were more specialized in feeding site than in
breadth of social behavior, tending to be more social in shallower water.

Habitat relationships. — The PC analysis of the 14 non-social factors depicted
the habitat space occupied by the whole waterfowl community and delineated
the realized habitat niches ( Hutchinson 1957, 1965) exhibited by the various
species within this space. The first principal component ( PC-I I showed that
water depth at feeding site, vegetational percentages, calcium, and conductivity
were the factors contributing to the most variability for waterfowl as a group
(Table 1) ; these were the same factors identified by PC-II in the preceding

101

TIIK WILSON BULLETIN • Vol. 90, Nu. 1, March 1978

INCREASING WATER DEPTH AND DEEP-WATER PLANTS ►

INCREASING SHALLOW-WATER PLANTS AND PRODUCTIVITY

Fig. 2. aterfowl ordination with 1% confidence ellipses based on species values and
means for scores of tlie first labscissa) and second (ordinate) principal components; only
vegetational, physical, and chemical environmental factors included in the analysis.

analysis and characterized feeding site habitats. Oxygen content and pH
were highly correlated with the second principal component ( PC-II ) and per-
tained to the nature of trophic activity in the water (Table 1). High pH and
water oxygen levels are associated with sites dominated by photosynthesis in
submerged plants. Sites dominated by organic decay are relatively low in
oxygen and pH.

Waterfowl habitat preferences are shown in Fig. 2 with 1% confidence
ellipses representing relative niche sizes. The sequence of species from left
to right on the abscissa ( PC-I ) is similar to the equivalent PC-H from top to
bottom on the ordinate in Fig. 1 (the slight differences being due to the
elimination of social factor effects for the principal components in Fig. 2 I .
With respect to PC-H ( Fig. 2) the Lesser Scaup occupied sites with the highest
pH and oxygen levels, the Mottled Duek and Blue-winged Teal at the opposite
extreme. Interesting relationships occurred such as the Green-winged Teal
favoring more photosynthesis dominated sites in shallow water compared to
the Mottled Duck, while the Blue-winged Teal preferred decay dominated
sites in deeper water. The Canvashack and Redhead had the smallest niche
sizes with respect to habitat axes, both teals the largest. The 2 principal com-
ponents described 35% of the total habitat variance.

Comparison of PC correlations for the random habitat samples (Table 1)
with those from separate analyses of each waterfowl species (for the latter PC

W^hite and James • WATERFOWL HABITAT USE

105

data consult White 1975) identified the species that responded directly to the
existing habitat contrasted to those that made special habitat adjustments.
The Mottled Duck, Green-winged Teal, Shoveler, American Wigeon, American
Coot, Redhead, Canvasback, and Lesser Scaup exhibited moderate to high cor-
relations for some or all of the factors that were highly correlated with PC-I
for the random habitat samples ( White 1975). Thus they responded directly
to the range in habitat conditions that was available in the lakes studied. The
other species, showing deviant PC correlations, selected special habitat condi-
tions from the common conditions existing there. Even those that responded
directly to the existing habitat were separated along the total habitat dine as
was evidenced by the existence of different species means for habitat factors
(White 1975). Principal components following PC-I differed progressively
more among species, and between species and the random samples. This
stressed the differing species specific habitat responses associated with the
decreasing variance of the later PC’s since essential requirements would tend
to be constantly present and thus less variable.

Environmental differences. — Relative positions of waterfowl niches shown
by PC analysis do not indicate whether species actually differ in responses, but
employing multivariate analysis of variance showed that differences were
significant (a = 0.05). The latter analysis differs from PC analysis in com-
puting new variables (canonical variables ) which are linear functions of the
original ones, but stressing those factors that effectively separate respective
populations ( Sokal and Rohlf 1969). An associated step-down analysis indi-
cated that water depth at feeding site and % emergent vegetation, both im-
portant factors in the first canonical variable, were statistically significant in
separating species. Vegetation height also was highly correlated with the first
canonical variable. Floating and or submerged vegetation and calcium con-
tent were highly correlated with the second canonical variable and thus were
important in characterizing the species environments, but were not statistically
significant in separating the species.

By plotting means of the first and second canonical variables for each spe-
cies (James 1971), an ordination showing maximum separation is obtained
(Fig. 3). The species ordering follows an environmental dine from shallow
water on the left to deep water on the right with associated decreasing % emer-
gent vegetation from left to right.

To determine which of the species overlapped in their requirements, the first
canonical variable scores for all samples of each species were subjected to a
common one-way analysis of variance with Duncan’s multiple range test.
Four distinct groups were significantly separated [a — 0.05) from all others
and each group was associated with a particular segment of the a(}uatic com-
munity (Fig. 4). Overlap in niche re(iuirements among species along the

106

rilE WILSON BULLETIN • VoL 90, No. 1, March 1978

z <
< ^
S CC
a. O

cr?

< 2
§ o
I -I
Q- <
LU O

American

Wigeon

Blue-winged

Teal

Pintail

Gadwall

Green-winged .
Teal

Northern

Shoveler

American-

Coot

Mottled

Duck

/•

Fulvous W.
Duck

Ruddy

Duck

Ring-necked

Duck

Redhead

Canvasback

Lesser

Scaup

INCREASING WATER DEPTH ►

INCREASING SHALLOW-WATER PLANTS

Ek;. 3. Waterfowl ordination using means of the first and second canonical variables
from multivariate analysis of variance, thus stressing factors that provide maximum separa-
tion among species.

dine is represented hy the thin horizontal lines beneath the heavy ordination
line.

Segment A (Fig. 4) represented very shallow water (1-30 cm) with
abundant emergent vegetation near lake shorelines. The Mottled Duck occu-
pied this part of the littoral zone and was never recorded elsewhere. Segment
R contained 3 species ( Blue-winged Teal, Green-winged Teal, and Northern
Shoveler ) and represented the part of the littoral zone having moderate
amounts of semi-aciuatic and aquatic vegetation with shallow to moderate
water depths ( 30-88 cm ) . Segment C had abundant aquatic vegetation, sparse
emergent vegetation, and moderate water depths (88-114 cm I. Six species
( Fulvous Whistling Duck, Pintail, Gadwall, American Wigeon, Ring-necked
Duck, and American Coot) occupied this region mainly, but some overlap is
seen between species in segments B and C. Thus the species in these subgroups
were not as exclusive in requirements as was the shoreline restricted Mottled
Duck. Segment D, representing open deep water (114-213 cm) with little
emergent vegetation habitat was fre(iuented by the Ruddy Duck, Redhead,
Canvasback, and Lesser Scaup.

3 he general trend of the species ordering along a community transition is
well illustrated in Fig. 5. I he species arrangement on the abscissa is in the
same order as in Figures 3 and 4, and the means for water depth at feeding
site, % emergent vegetation, and % floating and/or submerged vegetation are
plotted for each species (see White 1975 for tables giving mean values). The

White and James • WATERFOWL HABITAT USE

107

T—

— 1 T"

1 1

1 r

3 I

1 ^
1
1
1

B 1

C 1

INCREASING WATER DEPTH ►

INCREASING SHALLOW-WATER PLANTS

Fig. 4. Ordering of waterfowl determined by mean values of the first canonical variable
(scale of values below heavy line). Each horizontal thin line underscores species subgroups
that overlap in environmental characteristics. Capital letters (A, B, C, D) designate sub-
groups that were significantly different from one another ( bracketing individual thin lines
that do not overlap).

distinctiveness of environmental preferences ( Weller 1975 ) for each subgroup
is evident. Notice that puddle ducks are found in shallow to moderate water
depths (Weller 1975), diving ducks in deep water, and the 2 groups are at
opposite ends of the dine. The most species, 9 in 2 subgroups ( B and C), are
rather closely packed in the middle of the sequence (Fig. 4) where water
depths are moderate (88-114 cm ) and truly aquatic vegetation is greatest
( Fig. 5 ) .

The pattern of waterfowl preferences vs. habitat availability is shown by
plotting the PC scores obtained from analyzing only the 60 random samples
and establishing a 95% confidence ellipse based on these samples ( Fig. 6 ) .
This represents most of the available habitat space at the study area on the
Welder Foundation grounds (social factors not included in random samples).
Increasing water depth and distance to shore from left to right on the abscissa
(Fig. 6) and amount of emergent vegetation increasing in the opposite direc-
tion were highly correlated with PC-I (Table 1). Aiiuatic vegetation de-
creasing upward on the ordinate was highly correlated with PC-II. Together
PC-I and PC-II accounted for 41% of the total variance. Directional cosines
from the random samples PC analysis were used as weights to generate cor-
responding PC scores for each species. The means of these scores for each
species plotted in Fig. 6 show that the species were clumped near the center of
the available habitat space.

1()‘> THE WILSON lUJLLETIN • Vol. 90, No. 1, March 1978

Fig. 5. Waterfowl environments rejiresented l)y mean values for 3 environmental
factors: water depth at feeding sites, % emergent vegetation, and % floating and/or sub-
merged vegetation.

DISCUSSION

Recent multivariate studies of birds in various habitat types indicate that
species may lie arranged horizontally as well as vertically in response to
vegetational characteristics. Grassland birds were distributed vertically in the
tall vegetation by differences in feeding height and horizontally by differences
in habitat preferences (Cody 1968). Forest birds were distributed vertically
and horizontally along a continuum from forest edge to mature forest ( James
1971 ) and old-field birds were scattered along a dine in shrubbiness habitat
( Posey 1974 ) . Our findings show that waterfowl too were distributed along an
environmental dine, but an aquatic one rather than terrestrial. Social char-
acteristics proved diagnostic too.

Despite much overlap in groups of waterfowl species, each species occupied
a definite position with respect to sets of environmental factors ranging from
very shallow water w ith abundant emergent vegetation to open deep water with
little emergent vegetation of any kind (Figs. 3, 4, and 5). Water depth at
feeding site and % emergent vegetation were the 2 factors that were significant

White and James • WATERFOWL HABITAT USE

109

Fig. 6, Limits of the existing habitat space represented by the 95% confidence ellipse
circumscribing the principal component values for the individual random samples (open
circles). Mean values for the waterfowl species are superimposed (closed circles).

(a = 0.05 ) in separating species. Species richness and density were concen-
trated where aquatic vegetation was most prevalent (Figs. 4 and 5). Un-
doubtedly factors not measured in this study, such as food types (Weller
1972), feeding behavior ( Lack 1971 ), and other aspects of food selection also
are important in separating waterfowl environments, as is evidenced in other
birds ( Betts 1955, Root 1969, Shugart and Patten 1972 j .

Certainly the niche requirements for each species will shift from season to
season (Wiens 1969) and care must be taken not to generalize for waterfowl
wintering grounds as a whole. In fact it would be difficult, if not impossible,
to determine the niche of a species in its entirety. However, the use of multi-
variate statistical methods provides important progress toward this end. These
procedures produced a representative characterization and interpretation of the
ecological niches of wintering waterfowl. Further, application of these or simi-
lar techniiiues may he useful in wetland management programs. In so doing it
would be difficult to manage for or against particular species within a sub-
group, such as within the subgroups identified in Fig. 4, because of broad over-
lap in habitat use by the grouped species.

THE WILSON HUI.LETIN • Vol. 90, No. 1, March 1978

SUMMARY

A compurativt* study of tlu* (“nvironmental rfdationships among 14 species of wintering
waterfowl was conducted at the Welder Wildlife Foundation, San Patricio County, near
Sinton, Texas during the fall and early winter of 1973. Measurements of 20 environ-
mental factors (social, vegetational, physical, and chemicals were subjected to multivariate
statistical methods to determine certain niche characteristics and environmental relation-
ships of waterfowl wintering in the a([uatic community.

Each waterfowl species occupied a uni(jue realized niche by responding to distinct
combinations of environmental factors identified by principal component analysis. One
percent confidence ellipses circumscribing the mean scores plotted for the first and
second principal components gave an indication of relative niche width for each species.
'File waterfowl environments were significantly different interspecifically and water depth
at feeding site and % emergent vegetation were most important in the separation. This
was shown by subjecting the transformed data to multivariate analysis of variance with an
associated step-down procedure. The species were distributed along a community dine
extending from shallow water with abundant emergent vegetation to open deep water with
little emergent vegetation of any kind. Four waterfowl subgroups were significantly
separated along the dine, as indicated by one-way analysis of variance with Duncan’s
multiple range test. Clumping of the bird species toward the middle of the available
habitat hyperspace was shown in a plot of the principal component scores for the random
samples and individual species.

Naturally occurring relationships among waterfowl were clarified using principal com-
ponent analysis and related multivariate procedures. These techniques may prove useful
in w^etland management for particular groups of waterfowl based on habitat preferences.

ACKNOWLEDGMENTS

This research was supported by the Rob and Bessie Welder Wildlife Foundation, Sinton,
Texas ( Vi elder Contribution No. 167). The Arkansas Audubon Society Trust Fund pro-
vided a greatly appreciated research grant. Additional thanks are extended to James E.
Dunn, Alan F. Posey, Tim Mantooth, and Joel Carver for their help with statistical
analysis. E. Bolen, L. Fredrickson, and L. Stickel provided critical reviews of the manu-
script.

Also, sincere appreciation and gratitude is extended to Betty Jean bite for her as-
sistance in data collecting and laboratory analysis, and for her constant encouragement to
the senior author.

LITERATURE CITED

Andrews, 1). V., R. Gnanadesikan, and J, F. Warner. 1971, Transformation of multi-
variate data. Biometrics 27 :825-840.

Bargman, R. 1962. Representative ordering and selection of variables. Coop, Res.

Proj. No. 1132, Final Rpt., U. S. Office Health, Educ. Welfare.

Bellrose, F. C. 1976. Ducks, geese, and swans of North America. Stackpole, Harris-
burg, Pa.

Bennett, L. J. 1938. The Blue-winged Teal. Iowa State Univ. Press, Ames.

Betts, M. M. 1955. The food of titmice in oak woodland. J. Anim. Ecol. 24:282-323.
Box, C. E. P. AND 1). R. Cox. 1964. An analysis of transformations. J. Royal Stat. Soc.
26:211-252.

W hite and James • WATERFOWL HABITAT USE

111

Cody, M. L. 1968. On the methods of resource division in grassland bird communities.
Am. Nat. 102:107-147.

Cooley, W'. W'. and P. R. Lohnes. 1971. Multivariate data analysis. Wiley & Sons,
New ^ ork.

Hutchinson, G. E. 1957. Concluding remarks. Cold Spring Harlior Symp. Quant. Biol.
22:415-427.

. 1965. The ecological theater and the evolutionary play. Yale Univ. Press, New

Haven.

James, F. C. 1971. Ordinations of habitat relationships among breeding birds. Wilson
Bull. 83:215-236.

Lack, U. 1971. Ecological isolation in liirds. Harvard Univ. Press, Cambridge, Mass.

MacArthur, R. H. 1958. Population ecology of some warlilers of northeastern conifer-
ous forests. Ecology 39:599-619.

Morrison, D. F. 1967. Multivariate statistical methods. McGraw-Hill, New Tork.

Orians, G. H. 1966. Food of nestling Yellow-headed Blackbirds, Cariboo Parklands,
British Columbia. Condor 68:321-337.

Posey, A. F. 1974. Vegetational habitats of breeding birds in Ozark shrubby old fields.
Ph.U. thesis, Univ. Arkansas, Fayetteville.

Ralph, C. P. 1975. Life style of Coccyzus pumilus, a tropical cuckoo. Condor 77:60-72.

Root, R. B. 1967. Niche exploitation by Blue-gray Gnatcatchers. Ecol. Monogr. 37:317-
350.

Russell-Hunter, W\ U. 1970. Aquatic productivity. Macmillan, London.

Shugart, H. H. and B. C. Patten. 1972. Niche quantification and the concept of niche
pattern. Pp. 283-327 in Systems analysis and simulation in ecology, Yol. II ( B. C.
Patten, ed.). Academic Press, N. Y.

SiNCOCK, J. L. 1963 (1962). Estimating consumption of food by wintering waterfowl
populations. Proc. 16th Annu. Conf. Southeast. Assoc. Game Fish Comm. 16:217-221.

Singleton, J. R. 1968. Texas’ mistaken Mallards. Texas Parks and W'ildl. 26:8-11.

SoKAL, R. R. AND F. J. Rohlf. 1969. Biometry. W. H. Freeman, San Francisco.

Steel, R. D. G. and J. H. Torrie. 1960. Principles and procedures of statistics with
special reference to the biological sciences. McGraw-Hill, New ^ork.

Weller, M. W. 1972. Ecological studies of Falkland Islands’ waterfowl. Wildfowl 23:
25-44.

. 1975. Habitat selection by waterfowl of Argentine Isla Grande. Wilson Bull.

87:83-90.

White, D. H. 1975. Environments of freshwater feeding sites of waterfowl in autumn
on the Welder Wildlife Refuge in southern Texas. Ph.D. thesis, Univ. Ark., Fayette-
ville.

W HiTMORE, R. C. 1975. Habitat ordination of passerine birds of the Virgin River Valley,
southwestern Utah. Wilson Bull. 87:65-74.

WTens, j. a. 1969. An approach to the study of ecological relationships among grass-
land birds. Ornithol. Monogr. No. 8.

WiNKWORTH, R. E. AND 1). W. GooDALL. 1962. A crosswise sighting tube for point
quadrat analysis. Ecology 43:342-343.

DEPT. OF ZOOLOGY, UMV. OF ARKANSAS, FAYETTEVILLE 72701 ( PRESENT ADDRESS,
DH\V: U. S. FISH AND WILDLIFE SERVICE, PATUXENT WILDLIFE RESEARCH

CENTER, GULF COAST FIELD STATION, P. O. BOX 2506, VICTORIA, TX 77901).
ACCEPTED 14 SEPT. 1976.

I{KPR()I)LCTI\ E SUCCESS AND FORAGING BEHAVIOR
OE THE OSIMIEV AT SEAHORSE KEY, FLORIDA

Rohkkt C. Szaro

I^evious accounts liave described the decline of reproductive success in
the Osprey i Fandion haliaetus) (Ames and Mersereau 1964, Reese 1970,
\\ iemeyer 1971 ). W ith these reports in mind, I studied the breeding biology
of Ospreys at Seahorse Key, Levy County, Florida. The principal aims of
this study were to investigate the breeding biology of an Osprey population
on an isolated marine island without serious pesticide problems ( Blus et al.
19741. Included in the study are features of Osprey breeding behavior and
feeding habits, and pesticide contents of Osprey eggs and food.

STUDY AREA AND AIETHODS

Seahorse Key is located approximately 4 km southwest of the town of Cedar Key
on Florida’s Gulf Coast, and is approximately 9 km from the mainland. Since 1936 when
Seahorse Key became incorporated into the Cedar Keys National Wildlife Refuge, the
l)ieeding Ospreys, White Ibises iEudocimus albus) , Brown Pelicans (Pelecanus oc-
cidentalis) , and other forms of wildlife have been protected.

Seahorse Key consists of 62.5 ha, 20.1 of which are mangrove swamp, 15.8 are in the
littoral zone below mean high tide, and 26.6 include the beaches and upland areas
(Wharton 1954). The key has large areas of mangrove swamp in which many mangroves
were killed by freezes in January 1960, 1963, and 1966. The mangrove swamp includes
principally black mangrove (Avicenna nitida) , (90-95% of the mangrove area ) with
some white mangrove i Laguncularia racemosa) on the northern fringes and a rare
occurrence of red mangrove i Rhizophora mangle) scattered throughout the swamps
I Wharton 1954) . The mangrove swamps, found exclusively on the mainland side of
the island and lining its northern basins < Fig. 1), are in a period of regrowth, but each
year fewer and fewer of the dead mangroves are strong enough to support the weight
of an Osprey nest. The littoral zone has numerous shoal areas in the northern basins
and on the southern side of the island that are used for fishing by the Ospreys. In
addition, the east and west hanks on the southern side of the key are heavily used by
feeding Ospreys.

During the 1972 breeding season I checked each Osprey nest at least weekly, beginning
on 6 March 1972 and ending on 30 July 1972. Nests were examined either by climbing
or with the aid of a mirror on a long pole. Active nests were defined as those nests with
eggs or, in those nests not readily observable, by the presence of young at the nest.
Nest height from the ground or to sea level was measured with a steel tape and a mean
heiglit and standard deviation calculated.

Information on Osprey foraging behavior was olRained with binoculars and a stop-
watch. Each fisliing effort was timed from the moment an Osprey first was sighted until
it captured a fish or was lost from view. Fishing efficiency was determined by the
percent of fishing attempts resulting in fish capture.

'Ihree eggs, each containing a rotten hut well developed embryo, were taken from
Ospre\ nests after these eggs had (‘xceeded the normal incubation period by at least a

112

Szaro • OSPREY BREEDING SUCCESS

113

Fig. 1. Mangrove areas on Seahorse Key, Florida (modified from Wharton 1954) and
location of Osprey nests in 1972.

week. Along with the eggs, 3 fish, somewhat dehydrated, were also retrieved from the
same nests. Samples of fish muscle and homogenized egg contents were separately
ground with sodium sulfate. Subsequently each sample was extracted with petroleum
ether in a Soxhlet apparatus for 8 h, followed by acetonitrile partitioning and florisil
clean-up. Using the silica gel-pentane separation technique of Snyder and Reinert
(1971), the resulting sample was treated for the separation of polychlorinated biphenyls
(PCB's) from DDT and its metabolites, and dieldrin. Analysis for the DDT com-
ponents and dieldrin was performed on a Varian Aerograph model 600-D gas chromato-
graph using a column of OV 210 (6.4%) and OV 17 (1.6%) and operated isothermally
(oven and detector at 209°C). The PCB analysis was performed on a Varian Aerograph
model 2100 gas chromatograph using a similar column and operated with the oven at
206°C and the detector at 218°C.

RESULTS AND DISCUSSION

Reproduction. — During 1972 the 20 Osprey nests on Seahorse Key
ranged in height from 2 to 14 m ( x = 6.1 ± 3.2 m) with 85% of the nests on
the north side of the Key (Fig. 1). The major factor determining Osprey
nest height was the height of the dead black mangroves as 75% of the nests
were in mangroves. Sixteen of the nests were below 8 m in height. The four

Ill

I HK WILSON BULLETIN • Vul. 90, No. 1, March 1978

nesls above o.O m in height were in living oak [Quercus virgiiiiana) or
cedar {Sabina silicicuhi). Two of these nests were inactive hy 26 March
1972. dhree nests in the mangroves were inactive by 3 April 1972; the
remaining 15 nests were classified as being active.

Six closely observed nests had 3 eggs each with a hatching success of 44%.
1 hree of the resulting hatchlings were lost in early May when 2 nest sites
were destroyed by a storm. Eight of the 15 active nests produced a total of 11
fledglings.

Henny and Wight (1969 ) stated that each adult female in an Osprey popu-
lation must produce between 0.95 and 1.30 young per nest each year in order
to maintain a stable population. The 1972 production level at Seahorse Key
of 0.73 young per nesting female was below that needed for a stable popu-
lation. However, if not for the storm in May, production would have been
0.93 young per nesting female.

Foraging, — The Osprey preferred the shallow waters afforded by low tide
for fishing. The birds fed mainly during the middle-late ebbing to the
middle rising of the tide in the Cedar Keys area. They showed a definite
preference for the shoal areas consisting of the East and West Banks and
the shallow areas on the south side of the Key. The northern basins were
used for fishing but not as extensively as the fishing grounds on the south
side. The shallower waters in these areas before and after low tide were
likely favorite areas for fishing because of the large numbers of speckled
trout [Cynoscion nehulosus) and striped mullet {Mugil cephalus) that
freciuent them.

Despite their quick plunges. Ospreys required several strikes and many
minutes to catch a fish (Table 1). Overall, adult Ospreys were successful
in 58% of their fishing efforts. Ueoka and Koplin ( 1973 ) found that adult
Ospreys in northwestern California were successful in 82% of their fishing
efforts. At Seahorse Key, fish capture by adult Ospreys occurred on the first
attempt in 52% of the 23 successful fishing efforts. Of the 17 unsuccessful
fishing efforts, 7 (41%) involved Ospreys that flew back to the key without
securing fish. The remaining 10 (58%) unsuccessful fishing efforts in-
volved birds that were lost from view.

During the first week of flying, the young followed the adults to the shoal
areas and attempted to fish for themselves. Twice adults used food as a
method of coaxing the young to feed for themselves. An adult, carrying a
fish, flew around the young and finally dropped it in mid-air, making it
necessary for the young bird to dive quickly in order to catch it before it hit
the water. In this manner the young were forced to retrieve the fish from
the surface of the water until they finally started fishing by themselves.

Szaro • OSPREY BREEDING SUCCESS

115

Table 1

Analysis of Fishing Efficiency of Adult and Fledgling Ospreys

Adult

Young

No. fishing efforts

No. fish caught

Percent of fishing efforts successful

57.5

28.6

Attempts/catch

5.4 ± 6.7*

12.7 ± 7.2

Minutes/catch

38.3 It 21.7*

77.3 ± 32.8

Percent of attempts resulting in a catch

18.6

6.3

* Mean it standard deviation; Significant difference
using Student’s T Test.

between adults

and young at P — 0.05

This account is similar to that of Meinertzhagen (1954) for the luring of
young Ospreys to fish on their own.

Recently fledged Ospreys were less efficient at fishing than their parents.
The young birds made more attempts and required a longer period of time
to catch a fish than their more experienced parents (Table 1).

Ospreys at Seahorse Key fed primarily on speckled trout; of 103 food
items identified, 64% were speckled trout. Of the remaining fish caught,
27% were striped mullet, 8% were sea catfish {Galeichthys felis), and 2%
were ocellated flounder ( Amclopsetta quadrocellata) .

Eggshell thickness. — Seven samples of Osprey eggshells were obtained and
sent to D. W. Anderson for measurement. The shells ranged in thickness
from 0.38 to 0.53 mm with an average thickness of 0.46 ± 0.06 mm. Mem-
brane thickness was 0.12 ± 0.01 mm with extremes of 0.11 and 0.14 mm.
The calcite layer had an average thickness of 0.34 ± 0.06 mm with the
values ranging from 0.24 mm to 0.40 mm.

Anderson I pers. comm, in 1972 ) reports a “normal” value lor Osprey
eggshell thickness as 0.50 ± 0.01 mm in 20 eggs from Florida ( specific loca-
tion and dates not provided). The mean membrane thickness was 0.13 mm,
and the mean calcite layer thickness was 0.37 mm. Prior to 1947 average
shell thickness in Eastern United States Ospreys was 0.505 ± 0.004 mm
(Anderson and Hickey 1972). These data suggest that Osprey eggs at Sea-
horse Key were 8 to 9% thinner than “normal,” though interpretation of
shell thickness is made difficult by lack of knowledge concerning stage of
embryo development of eggs measured.

Pesticide analyses. — Three Osprey eggs and 3 fish samples taken from
active nests were analyzed on a lipid weight basis to determine organochlorine
pesticide and PCB residues. The eggs contained an average concentration
of 8.34 ± 1.45 ppm of p,p'-DDE, 3.55 ± 0.24 ppm of p,p'-DDD, 0.30 ± .02

116

THE WILSON BULLETIN • VuL 90, No. 1, March 1978

Table 2

DDT Compounds, Diki.dhin, and Bolychlokinatkd Bipiiknyls in Osprey Food Fishes

(IN PPM

LIPID basis;

Sample

p,p'-DDE

p,p'-DDD

p,p'-DDT

Dieldrin

PCB

Fish P

0.08

0.00

0.05

0.00

3.45

Fish 2*

1.82

0.00

4.54

o.co

227.30

Fish 3*

0.07

0.03

0.09

0.02

2.65

* Fish

1 = sea catfish, Fish

2 ocellated fl<

)iinder, Fish 3

= speckled trout.

ppm of p,p'-DDT, 0.26 ± 0.02 ppm of clieldrin, and 29.9 ± 6.7 ppm of
PCB ( Aroclor 1254 ) . The fish contained low levels of organochlorines and
only one, a sample of ocellated flounder contained high levels of PCB’s
( Table 2 ) .

Pesticide and PCB burdens in Osprey eggs and their food fishes have
been reported by several investigators ( Dustman et al. 1971, MacCarter et al.
1969, Stickel et al. 1965, Wiemeyer et al. 1975). DDE residues found in
Connecticut Osprey eggs were 254 ppm (lipid weight) (converted from
wet-weight by using the fact that fresh Osprey eggs contain up to 3.5%
lipid, Wiemeyer pers. comm.j, and in Maryland Ospreys DDE residues
were 69 ppm, (lipid basis). Residues in eggs from Seahorse Key were
much lower than these. Similarly, Connecticut Osprey eggs contained an
average of 17 ppm (lipid weight basis) of dieldrin as compared to only
.26 ppm dieldrin in Seahorse Key Osprey eggs. Dustman et al. (1971) re-
ported a median concentration of PCB’s of 15.9 ppm ( wet-weight basis )

( approximately 454 ppm on a lipid basis ) from eggs in Connecticut. Thus
the 3 Osprey eggs from Seahorse Key contained low pesticide and PCB
burdens.

Duke et al. ( 1970 ) report Aroclor 1254 from the water, sediment, and
biota of Escambia Bay, Florida. The Aroclor content in the water from
Escambia Bay was < 1 pph. Their study showed biological magnification
of Aroclor 1254 in a food chain: sediment contained < 0.3-1. 7 ppm;

crustaceans contained 1. 0-7.0 ppm; fish contained 415-184 ppm. It
is of interest to note that the PCB concentrations of their fish ( including
species examined in the present study ) were similar to those in the fish
taken from Osprey nests at Seahorse Key (Table 2).

\ here is, finally, no concrete evidence to suggest that any of these
chlorinated hydrocarbon residues played any significant role in influencing
hatching success of these birds especially since eggshell thicknesses of these
birds were near “normal.”

Szaro • OSPREY BREEDING SUCCESS

117

SUMMARY

During 1972 there were 15 active Osprey nests on the 62.5 ha of Seashorse Key,
Elorida. The Osprey population on this key produced 0.73 young per nesting female
which is l)elow the 0.95 to 1.30 young per nesting female necessary to maintain a
stable population. Adult Ospreys were successful in 18.6% of their fishing attempts.
The fishing technique of the Osprey is at least a partially learned behavior: adults

required only 5.4 attempts per catch and 38.3 min per catch, but the young required
12.6 attempts and 77.3 min per catch.

Pesticide analyses of 3 osprey eggs indicated low levels of organochlorines and
PCB's. Eggshells (nr=7) were approximately 9% thinner than shells collected prior
to 1947.

ACKNOWLEDGMENTS

I thank David W. Johnston for his generous help throughout my study and Pierce
Brcdkorb, Carmine Lanciani, John Ogden, and Stan Wiemeyer for their review of my
manuscript. I am grateful to Edward Collinsworth for granting me a permit to work
on Seahorse Key. A. I). Folks assisted me in travelling hack and forth to Seahorse Key
and Frank Mature provided living quarters and an outboard motor boat at Seahorse
Key. Research support was provided by the University of Florida, Department of
Zoology and from an Estuarine Ecology Grant through the Division of Biological
Sciences. I am grateful to Daniel Anderson for measuring my eggshell samples and
Ronald Bull for his assistance in the pesticide analyses.

LITERATURE CITED

Anderson, I). W. and J. J. Hickey. 1972. Eggshell changes in certain North American
birds. Proc. Int. Ornithol. Congr. 15:514-540.

Ames, P. L. and G. S. Mersereau. 1964. Some factors in the decline of the Osprey-
in Connecticut. Auk 81:173-185.

Blus, L. j., a. a. Belisle, and R. M. Prouty. 1974. Relations of the Brown Pelican
to certain environmental pollutants. Pestic. Monit. J. 7:181-194.

Duke, T. W., J. I. Lowe, and A. J. Wilson, Jr. 1970. A polychlorinated biphenyl
(Aroclor 1254) in the w-ater, sediment, and biota of Escambia Bay, Florida. Bull.
Env. Contain. Tox., 5:171-180.

Dustman, E. H., L. F. Stickel, L. J. Blus, W\ L. Reiciiel, and S. N. Wiemeyer. 1971.
The occurrence and significance of polychlorinated biphenyls in the environment.
Trans. North Am. Wildl. Nat. Res. Conf. 36:118-122.

Henny, C. j. and H. M. Wight. 1969. An endangered Osprey population: estimates
of mortality and production. Auk 86:188-198.

MacCarter, I). L., J. R. Koplin, and 1). S. MacCarter. 1969. Pesticides and re-
productive failure in Ospreys. Trans. Calif. Nevada Section of Wildl. Soc., 16:18-24.
Meineltzhagen, R. 1954. The education of young Ospreys. Ibis 96:153-155.

Reese, J. (i. 1970. Reproduction in a Chesapeake Bay Osprey population. Auk

87:747-759.

Synder, I). AND R. Reinert. 1971. Rapid separation of polychlorinated biphenyls from
DDT and its analogues on silica gel. Bull. Env. Contain. Tox., 6:385.

Stickel, L. F., F. C. Schmid, W. L. Reiciiel, and P. L. Ames. 1965. Ospreys in

THE WII.SON lUJLLETIN • VoL 90, No. 1, March 1978

IIH

Conm'cticul and Maryland. In PTfncts of I’esticides on Eisli and Wildlife. Fish
and Wildlife Serviee Cire., 226:4-6.

Ukoka, M. L., and .). K. Koplin. 1973. Foraging behavior of Ospreys in north-
western California. Raptor Res. 7:32-38.

Wii AHTON, C. H. 1954. The ecology of the (iottonmouths i Agkistrodon piscivorus
piscivoriis Lacepede) of Seahorse Key, Florida. Fh.D. thesis, Univ. of Florida,
Cainesville.

WiKMKYKH, S. N. 1971. Reproductive success of Potomac River Ospreys — 1970. Chesa-
j)eake Sci. 12:278-280.

WiKMEYER, S. N., P. R. Spitzeh, W. C. Krantz, T. G. Lamont, and E. Cromartie.
1975. Effects of environmental pollutants on Connecticut and Maryland Ospreys.
J. Wildl. Manage. 39:124-139.

DEPT. ZOOLOGY, UNIV. OF FLORIDA, GAINESVILLE 32611 (PRESENT AD-
DRESS: PATUXENT WILDLIFE RESEARCH CENTER, LAUREL, MARYLAND
20811). ACCEPTED 18 JAN. 1977.

REQUEST EOR ASSISTANCE

Purple Martin color-marking. — A large scale continent-wide Purple Martin color-mark-
ing project was initiated in 1977. Observers are asked to look for and report any color-
marked (plastic leg bands and or wing tags) Purple Martins. Please record the color
of the hands or wing tags, which leg they are on, age and or sex (if either is known),
where and when observed, and whether the bird was in a roost, staging flock, migratory-
flock, or at a nest site (scouting or nesting?). We are especially interested in the move-
ments of young birds and their return to the parent colony or nearby colonies. All re-
ports will he acknowledged and should be sent to Ms. Kathleen Klimkiewicz, Bird Band-
ing Laboratory, Laurel, Maryland 20811.

GENERAL NOTES

Changing: avian community structure during early post-fire succession in the
Sierra Nevada. — In August 1960 an intense fire consumed over 15,800 ha of pine-fir
forest in the northern Sierra Nevada near Truckee, Nevada Co., California. Establish-
ment in 1965 of 2 permanent study plots led to a comparison of breeding bird popula-
tions in burned and adjacent unburned habitats between 1966 and 1968 ( Bock and
Lynch, Condor 72:182-189, 1970). In 1975 we had the opportunity to census these
areas after 7 years of further post-fire succession. The purpose of this note is to describe
bird species diversity and avian community structure as they changed between 1968 and
1975.

Study areas. — This work w^as conducted at the University of California’s Sagehen
Creek Field Station, located 19 km N of Truckee. Each study plot was 8.5 ha, gridded
with permanent steel fenceposts set at 30 m intervals. The unburned plot is a mature
pine-fir forest, dominant species being Jeffrey pine ( Pinus jeffreyi) and white fir (Abies
concolor) . The burned plot contains a few scattered mature trees spared by the fire
including some Pinus ponderosa as well as A. concolor and P. jeffreyi), and especially
brush species such as Ceanothus velutinus and Arctostaphylos patula. There are sub-
stantial numbers of young pine. Between 1968 and 1975 there was a marked decrease in
standing dead timber and an increase in brush. For more detailed descriptions of the
vegetation, see Bock and Lynch ( 1970) and Bock et al. < Proc. Tall Timbers Fire Ecol.
Conf. 14:195-200,1974).

Census methods. — The census technique used was the \^’illiams spot-mapping method
• Williams. Ecol. Monogr. 6:317-408, 1936). This approach to estimating absolute den-
sities involves the repeated location of breeding birds on a grid, with clusters of ob-
servations eventually revealing the presence and territory sizes of breeding pairs. A major
problem with this method appears to be that different interpretations can be given to a
particular data set by different individuals (Best, Auk 92:452-460, 1975). Since all den-
sity estimations in this study were made by 1 person (CEB), and since accurate relative
abundances of species are sufficient for calculation of diversity and similarity indices in
any event, we feel that the technique was valid in this instance. The avifauna of the un-
burned forest should have changed little over the 7-year period. Similarity of census
results on the unburned plot in 1968 and 1975 (Table 1) supports our confidence in all
of the data collected.

Censuses were conducted from late May until early July, and varied from 1 to 3 h in
1968 and from 2 to 4 h in 1975. Numbers of censuses were as follows: 1968 unburned
plot — 15; 1975 unburned plot — 11; 1968 burned plot — 21; 1975 burned plot — 11. Bock
and Lynch ( 1970) include some detailed information on our particular approaches
to the spot-mapping method.

Results. — Table 1 is a summary of the census data for bird populations on the burned
vs. unburned study plots. Densities are expressed as pairs per 40.5 ha (100 acres) to
conform with most similar studies. Species richness, species diversity, and evenness all
were highest on the burned plot in 1968, lowest on the burn in 1975, and intermediate
on the unburned plot in both years. Although some of these differences are minor, diver-
sity on the burned plot was considerably higher in 1958 than in 1975.

Table 2 is a series of similarity indices comparing species densities on the plots in 19()8
and 1975. Two obvious trends emerge from these data. First, within-plot comparisons
(A, B) show that there was a much greater change over 7 years on the burned plot com-
pared to the unburned forest. Obviously this is a reflection of relatively rapid and

119

120

tup: WILSON BULLETIN • VoL 90, No. 1, March 1978

Table

Analysis of tmk Brkkding

Avifauna of the

Study Plots,

Expressed as

Pairs per

40.5 H A

Pairs/40.5 ha

Burned plot

Unburned plot

Species

1968

1975

1968

1975

Common Hioker

3.6

2.4

0.2

( Colaptes auratus)
Vellow-hellied Sapsucker
( Sphyrapicus vari us )
Williamson’s Sapsucker
(5. thyroideus)

2.4

1.2

0.2

White-headed Woodpecker
(Picoides alholarvatus)

1.2

0.2

Hairy Woodpecker
< P. villas us)

1.2

0.2

Three-toed Woodpecker
(P. arcticus)

3.6

1.2

Empidonax sp.

10.8

13.2

14.3

15.5

Western Wood Pewee

3.6

0.2

(Contopus sordididus)
Olive-sided Flycatcher
{Nuttalornis borealis)

0.2

1.2

Steller’s Jay
i Cyanocitta stelleri)

0.2

Mountain Chickadee

4.8

7.2

13.2

14.4

< Par us gambeli)
White-breasted Nuthatch
(Sitta carolinensis)

2.4

0.2

Red-breasted Nuthatch
( S. canadensis)

Pygmy Nuthatch
(S. pygmaea)

4.8

0.2

4.8

6.0

Brown Creeper
( Cert hi a j ami Haris )

2.4

4.8

3.6

House Wren
( T roglodytes aedon )

4.8

American Robin

7.2

4.8

0.2

{T Urdus migratorius)
Hermit Thrush
( Catharus gut tutus )
Mountain Bluebird
( Sialia currucoides )

15.5

10.8

2.4

7.2

March 1978 • GENERAL NOTES

121

Table 1 Continued

Pairs/40.5 ha

Burned plot

Unbumed plot

Species

1968

1975

1968

1975

Townsend’s Solitaire

1.2

i Myadestes townsendi)
Golden-crowned Kinglet

19.0

10.8

( Regal us satrapa)
Solitary Vireo
( Vireo sol it arias)

1.2

Nashville Warbler

3.6

( Vermivora ruficapilla)
Yellow Warbler
i Dendroica petechia)
Yellow-rumped Warbler
( D. coronata )

3.6

6.0

Western Tanager
( Piranga ladoviciana )

6.0

7.2

Cassin’s Finch
1 Carpodacus cassinii)

7.2

2.4

8.4

2.4

Pine Siskin

2.4“

( Carduelis pinus)
Red Crossbill
( Loxia curvirostra )
Green-tailed Towbee

2.4

9.6

1.2“

( Pipilo chlorura)
Dark-eyed Junco
ijunco hyemalis)

16.7

4.8

19.1

Chipping Sparrow

3.6

(Spizella passerina)
Brewer’s Sparrow
(S. breweri)

3.6

0.2

Fox Sparrow

2.4

29.9

2.4

0.2

iPasserella iliaca)

Totals

109.2

104.1

110.9

102.0

Species Richness

Species Diversity'’

4.07

3.41

3.57

3.50

Evenness‘S

.90

.79

.81

.80

® These 2 species probably were not breeding but were seen feeding on the unbumed plot in
1975 with such regularity that we have included them in the table at low densities. Their in-
clusion has little effect on overall density or species diversity measurements.

»//' =- V pj log., p. (see Peet, Annu. Rev. Ecol. Syst. 5:28.5-307, 1974).

i =1

c J - (see Peet 1974).

122

rilK WILSON BULLETIN • VoL 90, No. 1, March 1978

Taijle 2

SiMII.AKITIKS BkTWKKN TIIK BhKKDING AmFAUNAS OF Bl HNKI) AND U.NBIJKNED PlOTS IN 1968

AND 1975

C^oniparison Percent similarity®

Unburned (1968)

vs. unburned ( 1975)

Burned (1968) vs.

burned (1975)

Burned (1968) vs.

unburned (1968)

I).

Burned ( 1975) vs.

unburned (1975)

“Computed by S=(2W)/(« + b) (see Beals, Wilson Bull. 72:156-181, 1960).

dramatic successional events in the post-fire community. Second, and perhaps less ex-
pected, is the observation that breeding bird populations of the burned and unburned
forest were more similar in 1968 than they were in 1975 (Table 2, C and U). That is,
the 2 avian communities lost rather than gained similarity over the 7-year period.

Discussion. — In some respects the burned plot more closely resembled an unburned
forest in 1968 than it did in 1975. This fact may explain most of our findings, since it
is generally agreed that avian species diversity and community composition are tied to
habitat structure ( e.g., MacArthur, pp. 189-221, in Avian Biology Vol. I, Earner and
King eds.. Academic Press, N.Y., 1971). In 1968, 8 years after the fire, there were
numerous standing dead trees; there was much open ground, just as there is in an
unburmxl forest. By 1975 the burned plot was well on its way to being a dense brush-
field, with many fewer standing dead trees. Bird species characteristic of brush stands
in the area (Yellow^ Warbler, Green-tailed Towhee, Fox Sparrow) increased dramatically
on the burned study plot between 1968 and 1975. Open-ground foragers (American
Robin, Mountain Bluebird, Dark-eyed Junco) declined (Table 1), Six of 11 bole-nesting
species decreased on the burn between 1968 and 1975, while only 1 increased. This
species was the Mountain Chickadee, which appeared to forage extensively in the stands
of young regenerating pine on the burned plot.

Most, but not all, of the declining similarity between the burned and unburned plot
avifaunas was the result of population changes in the Dark-eyed Junco and Fox Sparrow.
These were the most abundant species on the burn in 1968 and 1975, respectively. Juncos
also were very common in the unhurned forest, nesting largely on the ground and forag-
ing on the forest floor. In 1968 juncos were similarly abundant on the burn (Table 1).
By 1975 the brushfields had closed off much of the burn and Fox Sparrows replaced
juncos as the most abundant breeding birds.

(Jdum (Ecology, 2nd ed., Holt, Rinehart, and Winston, New York, 1975:155) states
that “those species that are important in the pioneer stages are not likely to be important
in the climax.” Our findings do not support this generalization. It appears that bird
species populations changed on the burn in response to modifications of their individual
habitat reiiuirements. In some instances this resulted in convergence of the 2 avifaunas,
while in others it caused a decreased similarity between burned and unburned plot pop-
ulation levels. This is suggestive of the individualistic concept of community organization
proposed by ('deason (Am. Midi. Nat. 21:92-110, 1939).

Beaver ( Pb.D thesis, Univ. Calif., Berkeley, 1972) studied patterns of avian species
diversity in the .Sagehen Creek Basin, where this study was conducted. He compared
3 successional stages — brush, brush-conifer, and coniferous forest — and found that diver-

March 1978 • GENERAL NOTES

123

sity increased with habitat age. Examination of the data analyzed 1)\ Beaver (1972)
makes it apparent that only hy 1975 was our burned study plot beginning to resemble his
earliest or “brush” stage of succession, and to support a similar breeding avifauna. It
would appear that there exists an earlier “pre-brush” period of higher bird species di-
versity, followed by a decline when dead trees fall, when brush sj)ecies become dominant,
and when as a result there is much structural simplification of the post-fire community.

Vernon Hawthorne, Starker Leopold, and Marshall White generously made available
the facilities of the Sagehen Creek Field Station. This study was supported in part by
a grant from the University of Colorado Council on Research and Creative Work. —
Carl E. Bock, Martin Raphael, and Jane H. Bock, Dept, of Environmental, Population,
and Organismic Biology, Univ. of Colorado, Boulder 80309 (CEB, .IHB), and Dept, of
Forestry and Conservation, College of Natural Resources, Univ. of California, Berkeley
94720 (MR). Accepted 7 Dec. 1976.

Notes on the distrihution of birds in Sonora, Mexico. — Over the past several years
we have made observations of birds in northwestern Mexico that supplement published
distributional accounts. Many other individuals have contributed significant observations
in Sonora to us, so that we are able to elaborate on tbe status of 65 species of liirds, in-
cluding 20 species new for the state. Van Rossem ( Occas. Papers Mus. Zool. Louisiana
State Univ. 21:1-379, 1945) prepared the only major work on the birds of Sonora. Fried-
mann, et al. ( Pac. Coast Avif. 29, 1950) and Miller, et al. ( Pac. Coast Avif. 33, 1957)
are authors of the most recent check-lists covering the area and provide the basis for
deciding what distributional data should be included in this paper. At least 2 recent
works on Mexican birds (Alden, Finding the Birds in Western Mexico, Univ. Ariz. Press,
Tucson, 1969; Peterson and Chalif, A Field Guide to Mexican Birds, Houghton Mifflin
Co.. Boston, 1973) cited Sonora in ranges cf certain species for the first time but wdthout
locality or details; we document some of the reports on which these accounts are based.

Sonoran localities mentioned in the species accounts may be found on the map < Fig.
1). W'here specimens have been taken, we have indicated their present location as fol-
lows: Amadeo M. Rea Collection. Tucson, Az. (AMR), Delaware Museum of Natural
History, Greenville (DMNH), Museum of Vertebrate Zoology, Berkeley (MVZ), LTni-
versity of Arizona, Tucson ( UA ) , University of California at Los Angeles (UCLA). Im-
portant sight records are identified by the initials of the ol)server(s) as listed in the
acknowledgments. Our own observations are identifialile by our initials or by the use
of the word “we.”

Common Loon (Gavia immer) . — Zimmerman and Boettcher (Condor 69:527, 1967)
summarized Mexican records of this species and reported the first specimen from Mexico
(excluding Baja California). This loon is a common winter visitor along the coast of
Sonora with observations from 23 September to 28 April. Most March and Ajiril birds
are in breeding plumage. Two lurds found dead on the beach 4 December 1965 near
Puerto Pefiasco were preserved (UA); this date is 1 week later than the specimen re-
ported hy Zimmerman and Boettcher.

Red-tliroated Loon (Gavia stellata ) . — Van Rossem (op. cit. ) gave only 2 records from
the Gulf of California. Additional sightings are from 26 km SSE of La Libertad 27
November 1970 (SR), and Puerto Pehasco 15-18 March 1962 (SD), and 22 Fehruary
1975 (DS).

Horned (»rebe ( Podiceps auritus). — Earlier writers did not report this species from
Mexico at all, but Peterson and Chalif (op. cit.) referred to it as “casual” in Sonora.

121.

THE WILSON BULLETIN • VoL 90, No. 1, March 1978

It appears to be rare, though i)ossihly overlooked and specimens are lacking. Sightings
are from Puerto Pehasco (23 Octolier 1964, 3 birds WB, SR; 31 October 1970, 2 birds
RW), Punta Santa Rosa (29-30 November 1974, 2 birds SM). and Bahia Kino (21-23
February 1969, EC).

Northern Fulmar (Fidmunis glacialis) . — Two dark phase birds, carefully noted 24
March 1968 at Puerto Pehasco (William Bolte and RLC). constitute the only observa-
tion in Sonora.

(ireen Heron i Butorides striatus). — Van Rossem (op. cit.) had many summer
records but only 1 in winter. We found it in the mangroves at Punta Santa Rosa from
Sejjtember through February. It has also l)cen reported on the Sonoyta River in Decem-
ber (RLC).

March 1978 • GENERAL NOTES

125

Little Blue Heron {Florida caerulea). — The species was listed by Van Rossem (op.
cit.) as a summer visitor north to Guaymas with 1 winter record from Tiburon Island.
There are winter observations in mangroves at Punta Santa Rosa ( 16 December 1973,
DL. SR) and Bahia Kino (21-23 February 1970, WH, GB ) .

Reddish Egret { Dichromanassa rufescens). — Recorded by van Rossem (op. cit.)
south of Tepoca. This egret is a regular fall and winter visitor in small numbers along
the Gulf coast as far north as Puerto Pehasco (extreme dates, 7 Septemlier and 14 April).

Louisiana Heron i Hydranassa tricolor). — Van Rossem (op. cit.) listed it as an un-
common resident from Guaymas southward. Currently it is a common visitor along the
coast throughout the year. Observations of up to 6 birds together extend north to Puerto
Pehasco and are mostly between 7 September and 25 April.

Least Bittern ( Ixobrychus exilis ). — We took 2 specimens (UA), a male (wt. 64 g)
with whitish irides on 8 September 1973 and the other a female (81 g) with yellowish-
white irides on 17 September 1974, at Punta Santa Rosa. R. W. Dickerman had identified
them as /. e. pullus, the first from Sonora since the type series (van Rossem, Trans. San
Diego Soc. Nat. Hist. 6(15) :227-228, 1930). Both of our birds were captured in mist nets
set in mangroves i Avicennia germinans, Laguncularia racemosa) .

The birds did not have active gonads and we have no evidence that the species nests
at Punta Santa Rosa. We banded and released an additional adult at the same locality 18
August 1975.

W^hite Ibis i Eudocirruis albus). — Van Rossem (op. cit.) cited 1 record north of
Guaymas. Recent sightings are from Bahfa Kino (21-23 February 1969, EC), Punta Santa
Rosa (18 January 1974, DL, SR), and north of Puerto Pehasco (7 May 1975, SL).

Roseate Spoonbill iAjaia ajaja) . — Van Rossem’s (op. cit.) most northerly point of
record was Bahfa Kino (September). One was seen 26-27 June 1973 at Punta Santa
Rosa (JCL, DL). Inland, we saw 3 adults with Snowy Egrets iEgretta thula) on flooded
fields at Bamori on 9 September 1973.

Brant iBranta bernicla nigricans). — Grinnell ( Univ. Calif. Publ. Zool. 32, 1928)
recorded the Black Brant as a common winter visitor along the Pacific coast of Baja
California, but mentioned only a single “not quite conclusive” record for Gulf waters.
There are no published records from elsewhere in western Mexico. In recent years there
has evidently been a substantial shift in the wintering grounds of this subspecies. Since
1970, flocks of brant have been seen regularly in winter ( November-February ) at Bahfa
Kino and Punta Santa Rosa. Usually these groups number 150-450 birds, but about 2000
were observed in November (GM). Single birds have also been seen elsewhere; one 8
June 1973 at La Libertad (JJL) and 1 or 2 on 23-31 March 1974 at Puerto Pehasco (JW,
RW).

Common Goldeneye i Bucephala clangula) . — Booth (Condor 55:160, 1953) reported
on a pair seen at Guaymas in February 1951. The only additional records are sightings
by different observers from 9 November to 30 March from Puerto Pehasco, Bahfa Kino,
and Guaymas. No more than 4 birds were seen on any occasion.

Oldsquaw' {Clangula hyemalis). — Huey (Trans. San Diego Soc. Nat. Hist. 5:11-40,
1927) observed 1 bird in the Gulf near San Felipe. B.C. on 1 April 1926, Imt Friedmann
et al. (op. cit.) did not list it in Mexico. Peterson and Chalif (op. cit.) listed it as ac-
cidental in the Gulf of California. A male was seen on 28 November 1970 at Tepoca ( WC,
DL, SR).

White-winged Scoter [Melanitta deglandi). — One was seen at Puerto Pehasco 1 April
1973 ( SD et al.) and reported by Monson (Am. Birds 27:803-806, 1973). Another was
seen at the same locality 22-23 February 1975 (DS). Grinnell (op. cit.) listed it from

126

THE WILSON BULLETIN • VoL 90, No. 1, March 1978

tlie (rulf sitle of Baja California and Friedmann et al. (op. cit.) reported it only from
Baja California.

Ho iifjli-lefijsed Hawk iButeu lagopus). — Only Peterson and Chalif <op. cit.) and
Davis (A Field (iuide to the Birds of Mexico and Central America, Univ. of Texas Press,
Austin, 1972) list it, as a rare or occasional straggler to northern Mexico. Two adult
light-j)hase Birds were noted 6 January 1973, 21 km south of Nogales at an elevation of
about 1100 m (PA). .Another, also close to the border, was seen 16 February 1973 (W'R),
less than 2 km north of Cananea ( Monson, .Am. Birds 27:646-650, 1973).

Ferriigtiiums Hawk i Buteo regalis). — This hawk has been seen frequently at Puerto
Penasco with observations from 23 October to 2 March in all months except February
(P.A, WB, RL(J, SM, I)S, JW, RW), hut is unreported in the literature for northwestern
.Mexico.

Sandhill ( >ane (Grus canadensis) . — Van Rossem (op. cit.) reported it as formerly
abundant in the Colorado River delta and Rio Mayo valley hut indicated that numbers were
greatly reduced at the time of his writing. The only recent observations were about 200
seen 19 January 1974, in an agricultural area 60 km WSW of Hermosillo by R. Craig
and 2 on 13 October 1971 at a dam on the San Pedro River near Cananea ( SS, ES).

Virg:inia Rail i Rail us limicola) . — A specimen (wt. 55 g; U.A) taken in the mangroves
at Punta Santa Rosa 17 September 1974 was in fresh plumage. Cunningham saw 1 bird 23
December 1968 on the Sonoyta River close to the Arizona border. Van Rossem (op. cit.)
cited only 1 Sonoran specimen but considered it possibly resident in extreme northern
Sonora, as did Friedmann (op. cit.).

Black Oystercatcher i Haeinatopus bachmani). — Three were found in March and 1 in
May 1970 at Puerto Penasco (SD). Friedmann (op. cit.) reported it only from the Pa-
cific side of Baja California.

Piping: Plover i Charadriiis melodus). — Five of these plovers were observed at Puerto
Penasco 17 .April 1965 (SR), and another was noted (JW, RW) there 27 November 1971
(-Monson, Am. Birds 26:100-104, 1972). Four individuals were watched at Puerto Penasco
on 28 February 1972 ( PJ, RLT, RW ) and reported by Monson (Am. Birds 26:638-642,
1972). The species is not otherwise known from western Mexico.

Mountain Plover iCharadrius montanus). — Van Rossem (op. cit.) cited 3 Sonoran
records, only 1 subsequent to 1900. .At Cerro Colorado, in the Sierra Pinacate, 3 groups
totalling 45 birds were observed 22 November 1973 ( DL. SR), in a flat shrubless area
dominated by bare ground and supporting only scattered small plants of Cryptantha sp.,
iJalea (probably mollis or neomexicana) , Coldenia Palmeri and Euphorbia polycarpa.

Upland Sandpiper i Bartramia longicauda) . — One seen in grasslands about 16 km S
of Benjamin Hill 31 July and 8-9 August 1969 (RET) constitutes the only observation of
the species in Sonora.

Wandering: Tattler ( Heteroscelus incanus). — Van Rossem (op. cit.) noted only 1
individual (seen 21 April 1930) in Sonora. Friedmann (op. cit.) reported it from Sonora
without comment. The species was seen 22 April 1966 (DL) and 23 April 1974 (JW,
RW) at Puerto Penasco.

Red Knot (Calidris canutus) . — .Although cited few times in the literature, this species
is evidently a regular spring and fall migrant and winter visitor along the coast, sometimes
in numbers (over 100 at Puerto Penasco, 13 April 1973, SL). Records from Puerto
Penasco and Bahia Kino extend from 25 September to 28 .April.

Stilt Santipiper i Micropalarna himantopus) . — We saw 1 on flooded fields at Bamori
9 .September 1973. There are no ])ublished records of this species in Sonora.

Poniarine Ja«‘g:er ( Stercorarius pornarinus) . — The species is unreported from Sonoran

March 1978 • GENERAL NOTES

127

waters. On 11 April 1976, Arnold R. McGill saw 2, 1 harassing an Elegant Tern {Sterna
elegans), close to shore at Bahia Kino.

Parasitic Jaeger iStercorarius parasiticus) . — The only observation of this species in
Sonora is that of 1 light-phase adult seen 29 December 1974 at Puerto Pehasco (DS).

Glaucous-winged Gull (Larus glaucescens) . — Two records are cited from Sonora by
Friedmann (op. cit.). We know of 4 additional observations: single birds on 29 No-
vember 1970 at Punta Santa Rosa ( PG. DL), and at Bahia Kino 21-23 February 1969
(EC et ah), 21-23 February 1970 (WH et ah), and 28 April 1972 (GM).

Sabine’s Gull (Aema sahini) . — The only Sonoran reports are from Puerto Pehasco,

3 birds 5-6 October 1969 ( RW ) and an immature 13 October 1969 (REG).

Common Tern i Sterna hirundo). — Although unreported in the literature from

Sonora, it may be a common transient at Puerto Pehasco. Many flew south past there 14
July 1968 (DL) and 120 (including immatures) passed in 4.5 hours on 13 Octol)er 1969
(REG). One bird was noted 24 March 1968 (REG). There is one winter record, 3
January 1955 (RSC). A specimen (UA) was taken 25 September 1965. W'e know^ of only
one observation further south, a single bird at Punta Santa Rosa 11 October 1973 ( DL,
SR).

Least Tern ) Sterna albifrons) . — Reported by van Rossem (op. cit.) as a summer
visitor from Guaymas southward. There is a small nesting colony at Puerto Pehasco but
no specimens have been taken and the subspecies is hence uncertain. Birds arrive at the
end of March and are present in April ( GG, FT, JJL, DL), nesting in June (RW), and
absent by 13 July (DL). There are a number of records of migrating birds (April and
August-September) from coastal localities.

Black Tern {Chlidonias niger) . — Phillips and Amadon (Condor 54:163-168, 1952)
collected the only specimen of the Black Tern from Sonora near Caborca on 31 October
1948 but there are many observations between 13 July and 31 October from coastal lo-
calities. The only spring observation was one bird 28 April 1972 at Bahia Kino (GM).

Black Skimmer i Rynchops nigra). — This species has not previously been reported
north of Guaymas. One was seen at Puerto Pehasco on 24 April 1974 (JW, RW).

Long-eared Owl {Asia otus). — The species has been found in Sonora only on Ti-
buron Island in late December 1931 (van Rossem, op. cit.) and at Puerto Pehasco 23
October 1964 (WB) in a grove of widely spaced, small (3 m) mesquite (Prosopis) trees.

Buff-collared Nightjar i Caprimulgus ridgwayi) . — Even though this nightjar has been
recorded in both Arizona and New Mexico, there are no published records for northern or
central Sonora. We are aware of only 2 reports. Phillips and van Rossem found several
individuals calling and feeding on hillsides at El Gavilan on 4-5 May 1947 and collected
2 (DMNH, UCLA). They heard others on 5 May 8 km E of Mazocahui. On 9 July 1974,
several were heard (SA, TJ, SS) calling between Mazocahui and Moctezuma ( Alden and
Mills, Am. Birds 28:933-935, 1974).

Lucifer Hummingbird {Calothorax lucijer) . — We found female Lucifer Humming-
birds common in Arroyo Cajon Bonito (1280 m) on 16 May 1976 where they foraged
from the stream up the slopes. Four nests were all in the flood plain; 2 on small droop-
ing sycamore branches ( 3 and 4 m above the ground ) each contained 2 eggs as did another

4 m in a hackberry (Celtis). The fourth was under construction 2.5 m above the ground
in an ash. We saw no males, although 30 females w^ere counted in a 3 km transect of
the canyon. The only other hummingbirds present were many Broad-l)illed {Cynanthus
latirostris) , at least 2 Broad-tailed ( Selasphorus platycercus) . and 1 Black-chinned
{Archilochus alexandri) and nesting Blue-throated iLampornis clemenciae ) . On 13 July,
no Lucifer Hummingbirds were found in the canyon but 2 males and a female were noted

THE W ILSON HULEETIX • VoL 90, No. 1, March 1978

12H

(CM, SK ) on the dry upper slopes ( 1550 m) where most Agave had completed flowering.
'I’he canyon floor, where Lucifer Humminghirds dominated in May, now had many nesting
Broad-hilled, Black-chinned, and Violet-crowned humminghirds ( Amazilia verticalis) .
The species has not been reported previously from Sonora hut has recently been found
n«‘sting in Arizona < Monson, Am. Birds 27:803-806, 1973).

Allen’s Iliiininingliircl ^ Selasphorus sasin). — We collected a female (UA; rectrix 5
is 2.2 mm wide and tip of rectrix 2 is not emarginate) in the mangroves at Punta .Santa
Rosa 28 February 1971. Phillips (Condor 77:196-205, 1975) cited only 1 Sonoran speci-
men, a male taken near .Sonoyta, also in late February.

(ialliope Huimningbird (Stellula calliope). — The only known observation in a
coastal locality was of a male at Bahia Kino 28 April 1972 (CM) ; the species is not an
uncommon migrant in the eastern part of the state (van Rossem, op. cit.).

(ireeii Kingfisher i Ch/oroceryle americana) . — This kingfisher is widely distributed
in the interior of Sonora hut there are only 3 observations from coastal localities north of
Agiahampo. A male was collected at Tepoca 2 January 1932 (van Rossem, op. cit.). One
was seen (SD) at Puerto Pehasco on 1 April 1973 (Monson, Am. Birds 27:803-806,
1973). \^’e handed a male in the mangroves north of Chueca on 18 September 1974.

Vi’illiamson’s Sapsucker (Sphyrapicus thyroideus) . — A male was found ( ES, SS)
in the mountains (1800 m) near Cananea 13 October 1971 (Monson, Am. Birds 26:100-
104, 1972). Van Rossem (op. cit.) cited only 1 Sonoran record.

Tropical Kingbird ( ryr«nnir,s melancholicus) . — Both van Rossem (op. cit.) and Mil-
ler et al. (op. cit.) considered the northern coastal limit at Guaymas. The species is a
regular summer resident in mangroves north to Tepoca, with extreme dates of 28 April
and 19 August.

Scissor-tailed Flycatcher iMuscivora jorficata) . — We watched 1 individual on a
powerline with migrating kingbirds 10 October 1973 about 26 km south of Nogales, the
only observation in Sonora.

Hammond’s Flycatcher i Empidonax hammondii) . — The species is a regular transient
through the mountains of eastern Sonora (van Rossem, op. cit.) ; a male collected (UA)
in mangroves at Punta Santa Rosa on 7 September 1973 constitutes the only coastal record
(I)L, SR).

Horned Lark {Eremophila alpestris) . — W'e observed adults with recently fledged
young in the Salicornia dominated flats near the beach at Tepoca on 14 April 1973. W'e
saw a bird in the same locality on 16 July 1975 and 7 on 18 August 1975. No specimens
were taken and the subspecific affinities of these birds are unknowm; the locality is the
southernmost known one for the species in .'^onora. Previously the species had been found
only as far south as Altar, where non-hreeding birds were collected ( Phillips and Ama-
dou, op. cit.) .

Mangrove Swallow ^ Tachycineta alhilinea) . — The species is common along the coast
and a few kilometers inland as far north as Desemboque; northernmost records are from
8 km SSE of La Lihertad. It arrives in mid-March and nests in cardons ( Pachycereus
pringlei) in April and May. Most birds leave in mid-summer though there was an ob-
servation of 1 at San Carlos on 21 October 1975 ( GM ) . Several seen on 19 January 1974,
milling with Violet-green Swallows ^Tachycineta thalassina) over irrigated fields at El
Coyote ( I)L, !^R ) , may have been early spring arrivals.

(dark’s Nutcracker \Nucijraga cohunbiana) . — A clean skull and mandible found
near Puerto Penasco 16 December 1972 constitutes the second specimen (AMR) from
.Sonora. In autumn of 1972, nutcrackers wandered far from their usual range (even to

March 1978 • GENERAL NOTES

129

latitude 29° in central Baja California, 12 November 1972, RC and 25 November, DL,
SR). Miller et al. (op. cit.) cited a specimen from El Tigre, 14 June 1956.

Pinon Jay i Gymnorhinus cyanocephalus) . — Monson found 1 just across the Sonora
border in Chihuahua near the top of Sierra de San Luis on 1 May 1974. Eighty foraged
on the ground in mesquite grassland 6 km south of Sasahe on 11 November 1975 ( GM,

SR) , an extension of a flight that brought many to southern New^ Mexico and Arizona
(Witzeman, Hubbard and Kaufman, Am. Birds 30:105-110, 1976). There are no other
Sonoran records.

Brown Thrasher iToxostoma rufum). — Not listed by Miller et al. (op. cit.), hut
Peterson and Chalif (op. cit.) stated “casually wintering in northeast Mexico; also
Sonora.” It was seen at Guaymas in early December 1968 (PA). One individual at Maza-
tlan, Sinaloa, observed by Lamm on 12 April 1968, constitutes the only other western
Mexico observation to our knowledge.

Le Conte’s Thrasher iToxostoma lecontei) . — The most southerly published locality
(Stephens, Auk 2:225-231, 1885) is Puerto Lobos but we have found it in small numbers
south to Tepoca where it occurs in the Frankenia-Opuntia association.

Swainson’s Thrush i Catharus ustuJatus ) . — We netted 7 during the night of 17-18
September 1974 in mangroves at Punta Santa Rosa. Weights ranged from 27.0 to 34.5 g
and the heavier birds had considerable subcutaneous fat. Two kept as specimens (UA)
were identified as C. u. oedicus by A. R. Phillips. An intense thunderstorm a few kilo-
meters inland produced strong winds over the coast for a short time during the night.
Phillips and Amadon (op. cit.) speculated that Pacific Coast populations may migrate south
just inland from the Gulf of California. These birds, presumably forced down by winds,
support tbeir hypothesis. C. m. oedicus normally “does not stop in fall north of southern
Sonora” ( Phillips, Marshall, Monson, The Birds of Arizona, Univ. Ariz. Press, Tucson,
1964) and the species normally migrates through the eastern foothills (van Rossein, op.
cit.) .

Starling: iSturnus vulgaris). — Listed by Miller et al. (op. cit.) only from Tamaulipas.
In recent years it has become widespread in Sonora. Localities include Puerto Pehasco
( regularly at least since 1968, RLC, GM, RW ) ; 20 km S Nogales ( first observation Au-
gust 1968 ARP), El Coyote, 21 March 1974 ( DL, SR) ; Los Hoyas, 1 March 1974 ( DL,

SS) ; and near Navojoa investigating nesting sites, 5 April 1974 ( SS, DL).

Hutton’s Vireo iVireo huttoni). — The only records west of the 110th Meridian in

Sonora are from Sierra Carrizal (late October 1948; Phillips and Amadon, op. cit.).
Sierra del Humo (11 November, 1975, SR), and on the Sonoyta River (23 December
1968, RLC).

Prothonotary W'arbler i Protonotaria citrea). — A male was seen 26 April 1972 on
Rio Cucbujaqui, 11 km S of Alamos ( GM ) ; the species is not previously reported from
northwestern Mexico.

Virg:inia’s ^"arbler ( V ermivora virginiae) . — The only observation of the species from
a coastal area in Sonora was of 1 seen 15 April 1973, 16 km NE of El Coyote (SR),

Ovenbird iSeiurus aurocapillus) . — Although the species winters regularly in Sinaloa,
there has been only 1 Sonoran record, a singing male collected 21 June 1954 at Aconchi
(Marshall, Pac. Coast Avif, 32, 1957). We banded single individuals in mangroves north
of Chueca 21 June 1974 (wt. 16 g) and 28 May 1975 ( wt. 14.6 g). The birds were not
fat and had completely ossified skulls.

Hooded "^'arbler iWilsonia citrina) . — A male (UA) taken 30 January 1966 near
Alamos in a dense thicket adjacent to the Rio Cuchujaciui weighed 9.8 g. The species
was not cited from western Mexico by Miller et al. (op, cit.).

130

'I'HE WILSON nULLETIN • VoL 90, No. 1, March 1978

Riifoiis-capped Warbler ( Basileuterus rufifrons ). — On 13 September 1975, Danforth
et al. found a singing individual in Arroyo Cajon Bonito (Witzenian et al., op. cit.). Four
singing birds were present in tbe same area 14 May 1976 l)ut no mates were observed
( l)L, SR), The birds foraged on the north facing slope of the canyon (1280 m) in rock
crevices, in oaks and junipers on tbe slope, and in willows and sycamores of the flood
plain. Two males patrolled contiguous territories of about 2 ba, although they did not
sing or approach in response to playbacks of their own songs. None was found in mid-
July. The birds were about 180 km N of other localities of record ( Moctezuma and moun-
tains to SE) and only 6 km S of the United States.

Great-tailed Graekle iQuiscalus rrjexicanus) . — In recent years this species has ex-
tended its range virtually throughout western Sonora. It occurs along the Gulf coast at
many localites as far north as Puerto Pehasco, where it has occurred regularly at least
since 1965 (UL), Inland records include Sonoyta, Altar, Caborca, and Pitiquito.

Rose-breasted Grosbeak i Pheucticus ludovicianus) . — There are no published re-
ports for northwestern Mexico. One full-plumaged male was observed 26 May 1975 8
km SSE of La Libertad (SR) and S. C. Brown collected (MVZ) an immature male 24
September 1975 near Carbo, Sonora.

Indig;o Bunting: iPasserina cyanea) . — In a riparian Arizona Upland community in
Sonoran Desert scrub 65 km south of Sasabe, we noted a male singing constantly on 25
May 1975, while Varied Buntings held adjacent territories. He responded strongly to a
playback of his own song but did not react to recordings of Varied Bunting songs played
in what seemed to be his territory. The location was characterized by large flowering
mesquite (Prosopis juli flora), cat claw (Acacia Greggii), and buisache (Acacia Farnes-
iana ) . Sahuaros ( Cereus giganteus ) were beginning to bloom and although no grasses
or annuals were present, the vegetation present was green. On 15 July 1976 Doug Stotz
found a male singing in a riparian community dominated by mesquite and a few syca-
mores (elev. 1250 m) near Arroyo Cajon Bonito in northeastern Sonora. Varied Bun-
tings sang nearby.

Varied Bunting: (Passerina versicolor) . — Although reported only from eastern Sonora
by Miller et al. (op. cit.), the Varied Bunting is a summer nesting species in dense
riparian communities in the Arizona Upland Subdivision of the Sonoran Desert scrub at
least as far west as Caborca. Its occurrence in a given locality in any year seems to be
dependent upon adequate precipitation. We collected (UA) a female in the mangroves
near Chueca on 22 June 1974. The bird, moderately fat and weighing 12.2 g, possessed a
partially ossified skull and an ovary less than a mm in diameter. Its worn plumage is
comparable to that of females at least 1 year old of P. v. dickeyae. We know of no nesting
localities in the area.

Golden-crowned Sparrow' ( Zonotrichia atricapilla) . — Once recorded at Caborca (Mil-
ler et al,, op. cit.) and 3 seen at Puerto Pehasco 3 January 1965 (SR).

Wbite-throated Sparrow (Zonotrichia albicollis) . — On 2 November 1975 D. Danforth
found a single White-throated Sparrow in a dense thicket adjacent to the stream in Ar-
royo Cajon Bonito at an elevation of about 1250 m. The species occurs regularly in
southern Arizona, and this first observation in Sonora emphasizes the paucity of ob-
servers in northern Sonora.

Fox Sparrow (Passerella iliaca). — Cunningham observed 1 on tbe Sonoyta River 23
December 1968. There are no other Sonoran records.

A great number of individuals have extended their field activities to northwestern
Mexico. We are pleased to acknowledge the contributions of the many persons who

March 1978 • GENERAL NOTES

131

shared their observations with us. The following persons made observations which we
acknowledge in the text by their initials:

P. Alden (PA), S. Alden ( SA ) , G. T. Bancroft (GB), W, Buhner (WB), E. Chalif
(EC), R. Chapin (RC), W. C. Clark (WC), R. S. Crossin (RSC), R. L. Cunningham
(RLC), S. R. Demaree (SD), W. Harrison (WH), P. Gould ( PG ) , G. Gregg (GG), P.
Jeheher ( PJ ) , T. B. Johnson ( TJ ) . J. J. Levy (JJL), Seymour H. Levy (SL), J. C.
Luepke (JCL), G. S. Mills ( SM ) , G. Monson (GM), A. R. Phillips (AP), Mrs. Win-
hourne Ranney (WR), E. L. Smith (ES), S. Speich (SS), D. Stejskal (DS), F. Tainter
(FT), R. L. Todd (RLT), R. E. Tomlinson (RET), J. Witzeman (JW), R. A. Witzeman
(RW).

We thank E. Eisenmann, G. Monson, and A. R. Phillips for reading the manuscript
and offering suggestions. We also appreciate the help of R. Felger who identified some
plants for us and A. R. Phillips, R. Dickerman, and A. M. Rea for identification of some
bird subspecies. Permission to collect in Mexico was provided by the Direccion General
de Fauna Silvestre and we are appreciative of this courtesy. — Stei’Hen M. Russell, Dept,
of Ecology and Evolutionary Biology, Univ. of Arizona, Tucson 85721; and Donald W'.
Lamm, 6722 East Nasurnpta Dr., Tucson, AZ 85715. Accepted 7 Dec. 1976.

Eg:g carrying by Wood Duck. — On 4 April 1976, we observed a hen Wood Duck
i Aix sponsa) carrying an egg with her hill exit a W ood Duck nest box located in a
beaver pond approximately 1.6 km west of Indian Mound, East Baton Rouge Parish,
Louisiana. We could not determine whether the entire egg was being carried between
the upper and lower mandibles or if the lower mandible was inserted into a hole in the
egg. Earlier that same morning, we had seen a Wood Duck flying low over the heaver
pond similarly carrying an egg-shaped object that it dropped into the water. Previously,
on 2 April 1976, this nest contained 15 eggs in their 26th day of incubation. The average
incubation period of Wood Ducks is 30 days (Bellrose, 111. Nat. Hist. Surv. Circ. 45, 1953).

On the afternoon prior to the egg-carrying observation, we discovered that several eggs
of this clutch had holes pecked in them. These holes may have been made by wood-
peckers. Cunningham ( Proc. S. E. Assoc. Game and Fish Comm. 22:145-155, 1968 ) has
observed Red-hellied Woodpeckers i Melanerpes carolinus) pecking Wood Duck eggs
without consuming them. He found that Common Flickers ( Colaptes auratus ) and Red-
headed Woodpeckers iMelanerpes erythrocephalus) are also Wood Duck nest predators.
All 3 woodpeckers were present at our study area. Shortly after w'e observed the duck re-
moving the egg from the nest box that contained the damaged eggs, we discovered that
only 7 eggs remained in the nest box. The nest was subsequently abandoned and the re-
maining eggs failed to hatch.

There have been previous reports of egg-carrying by waterfowl. Sowls ( Prairie Ducks,
Stackpole Co., Harrisburg, Pa. and Wildl. Manage. Inst., Washington, D.C., 1955:104-
108) observed hen Northern Shovelers (Anas clypeata) and hen Pintails (A. acuta)
carrying egg shells away from their nests that had had some of their eggs destroyed by
predators. He also was able to induce an egg-carrying response by placing egg shells on
top of eggs in active nests. Hochhaum ( The Canvashack on a Prairie Marsh, 2nd ed.,
Stackpole Co., Harrisburg, Pa. and Wildl. Manage. Inst., Washington, D.C., 1959:92)
witnessed a hen Northern Shoveler carrying an egg in the tip of her hill, and Lindsey
(Auk 63:433-492, 1946) observed a Mexican Duck (A. diazi) carrying an embryo with
remaining portions of a broken egg from its nest. He concluded that the egg-carrying he-

132

THE WILSON BULLETIN • VoL 90, No. 1, March 1978

havior he observed was a response to severe disturbance. Cunningham (op. cit.) reported
watching a hen Wood Duck fly from a nest with either a whole or the greater portion of
an egg in her hill. This same hen suhse(}uently brought off a brood of 9, although 6
eggs had disappeared.

The role of egg-carrying behavior by ducks is yet unclear. However, our observations
and the existing literature suggest that egg-carrying behavior is an adaptation that may
increase the probability of the successful incubation of undamaged eggs in partially-
destroyed clutches.— Kohkkt W. Stradkr, Kiciiard I)i GitJi.io, and Robert B. Hamilton.
School of Forestry and ildlife Management, Louisiana State Univ., Baton Rouge 70803.
Accepted 9 Dec. 1970.

Evidence of brood adoption by Ruffed Grouse. — Evidence that Ruffed Grouse
{Bonasa umbel lus) hens sometimes adopt chicks from other broods is scanty. Bump et al.
(The Ruffed Grouse, Life History, Propagation, Management, N.Y. State Cons. Dept., Al-
bany, 1947:293) reported a hen which had lost its clutch just before hatching that was
later seen with 4 chicks. Apparent instances of 2 broods traveling together have been
reported several times. Hungerford (Murrelet 34:35-40, 1953) noted a “brood” of 2
hens and 15 young. Chambers and Sharp (J. Wildl. Manage. 22:231-239, 1958) simul-
taneously captured 2 hens and 5 chicks of 2 age groups. On several other occasions they
reported flushing groups of grouse containing young of noticeable age differences. Bump
et al. (op. cit.: 293) reported occasional encounters of double broods with both hens
normally present but felt that these often represented simply chance meetings of broods.
They also believed that older broods may merge permanently at times, particularly if 1
hen is killed.

These reports consist primarily of chance sightings of unmarked individuals. The his-
tories of the individuals sighted and the number of true adoptions of chicks were largely
unknown. Here 1 present evidence that 2 radio-marked Ruffed Grouse hens adopted extra
chicks in 1972 at the Cedar Creek Natural History Area 48 km north of Minneapolis,
Minnesota. Brood hens were flushed at weekly intervals through early July and at ir-
regular intervals thereafter. Chicks were counted at these times. Typically, the number
of chicks seen gradually decreased as the season progressed. There were 2 exceptions.
Yearling hen 2239, which hatched 9 eggs, was seen with 9 chicks 20 days after hatch and
with 13 young 27 days after hatch. Three days later this hen was killed by a predator pre-
cluding additional data on her brood. Adult (22 months or older) hen 2235’s brood of 10
chicks was reduced to 4 by 30 days after hatch. However, on 4 occasions between 57 and
70 days after hatch this hen flushed with 10-15 chicks. Since home ranges of the 2 hens
were adjacent, Hen 2239’s chicks conceivably could have joined hen 2235’s brood. At no
time was a second adult seen with these broods.

My observations of both captive and wild Ruffed Grouse chicks indicate that by 28-35
days of age they are well-feathered, strong fliers, forage for themselves, are very adept at
spotting aerial predators, and possibly could survive on their own. However, broods nor-
mally remain together about 75-85 days (Godfrey and Marshall, J. Wildl. Manage. 33:
609-620, 1969). For this behavior to be maintained by natural selection there must be a
survival advantage afforded to chicks that remain with the hen for this length of time.
If this is the case, orphaned or lost chicks would benefit by joining another brood.

It is more difficult to explain what evolutionary advantage a hen obtains in accepting

March 1978 • GENERAL NOTES

133

another hen’s chicks. Kin selection has been used to explain seemingly altruistic behavior
in several species ( Brown, The Evolution of Behavior, W. W. Norton and Co., Inc., New
York, 1975:203). This concept seems unlikely to apply in the present circumstance since
the fall dispersal pattern of Ruffed Grouse (Godfrey and Marshall, op. cit., 1969) re-
duces the likelihood that adopted chicks would be closely related to the hen.

Clearly, there are situations (e.g., where food is limiting) when the presence of extra
chicks would be detrimental to a hen’s own young. Under these conditions a hen which
adopted chicks would be selected against. Whether Ruffed Grouse hens can recognize
their own chicks is not known. If brood adoption is disadvantageous, one would expect
selection pressures for hens to recognize their own chicks and exclude others.

Since extra chicks apparently are tolerated, the possibility remains that hens actually
benefit by accepting other chicks. After chicks are 2-5 weeks old. Ruffed Grouse eat a
wide variety of plant foods (Bump et ah, op. cit.: 850). Potential grouse food appeared
to be abundant at Cedar Creek during this time and likely minimized the disadvantages
of extra chicks in terms of competition for food. Where food is not limiting and where
the effects of extra chicks in terms of predator attraction are offset by the increased
probability of predator detection, a hen which adopts chicks or travels with another brood
may increase her relative fitness because any chick captured by a predator would be less
likely to be one of her own.

I am grateful to the personnel of the Cedar Creek Natural Histor>- Area and of the
University of Minnesota Bioelectronics Laboratory (UMBL) for their cooperation during
the study. I thank Richard A. Huempfner and Gary J. Erickson for assistance with the
field observations. Lewis W. Oring and George-Ann Maxson made critical comments on
the manuscript. This investigation was supported by the U.S. Atomic Energy Commission
(COO-1332-108) . — Stephen J. Maxson, Dept, of Ecology and Behavioral Biology, Univ. of
Minnesota, St. Paul 55101 (Present address: Dept, of Biology, Univ. of North Dakota,
Grand Forks 58202) . Accepted 7 Dec. 1976.

Marsh Hawks follow hunting: red fox. — At 11:00 on 11 January 1973, we ob-
served a red fox ( Vulpes fulva ) hunting among scattered clumps of dead herbaceous
vegetation in an otherwise heavily grazed pasture in northern Delaware County, Ohio.
Snow cover was net present. Although the fox had a severe case of mange, the animal's
behavior appeared normal. Its hunting behavior consisted of the typical canine search,
pause, and pounce sequence. Two Marsh Hawks [Circus cyaneus) were near; 1 hawk
circled at a low level over the hunting fox while the second bird perched on the ground at
approximately 9 m to one side of the fox. As the fox completed its hunting activities in
one clump of vegetation and moved to the next clump, 1 Marsh Hawk again perched on the
ground near the fox while the other bird circled overhead. When the fox had exhausted
the remaining huntable clumps in the general area and had proceeded off across the
pasture, the hawks again followed. The trio was then lost from view as the fox entered
an area of scattered woods at the end of the pasture. During the entire observation period
of approximately 15 min, prey w'as not taken by either predator.

Two hypotheses may be advanced to account for the behavior of the Marsh Hawks:
(1) the 2 hawks and the fox were involved in some form of cooperative feeding interac-
tion, and (2) the movement of a small- to medium-sized mammalian predator may
naturally elicit a following response among Marsh Hawks.

131.

THE WILSON BULLETIN • VoL 90, No. 1, March 1978

Cooperative f(‘(‘ding interactions involving two or more avian predators have been de-
scribed for a number of species (Christman, Condor 59:343, 1957; Parks and Bressler,
Auk 80:198, 1963; Meyerrieeks and Nellis, Wilson Bull. 79:236, 1967; Dusi, Auk 85:129,
1968; Emlen and Ambrose, Auk 87:164-165, 1970; Haversehmit, Wilson Bull. 82:99, 1970;
.Mueller et al.. Auk 89:190, 1972; Anderson, Wilson Bull. 86:462, 1974) ; however, only
one account of a cooperative feeding interaction l)ctween an avian predator and a mam-
malian predator is given in the literature. Welty (The Life of Birds, W. B. Saunders Co.,
Philadelphia, 1975:396) described a cooperative feeding interaction which involved a
Rough-legged Hawk (Buteo lagopus) that fed upon rodents dislodged by a hunting
Arctic fox ( Alopex lagopus).

The tendency for birds to follow mammalian predators in a situation which does not
involve nest site defense has been reported for several avian predator species. Berger
lAuk 73:288, 1956) gave an account of a Marsh Hawk pursuing a domestic cat (Felis
domestica) . A pair of Mountain Choughs ( Pyrrhocorax graculus) were reported by Lane
(Ibis 99:116, 1957) to follow a hunting stoat (Mustela erminea) . Holland (Br, Birds 67:
212-213, 1974) observed an attraction and following tendency among Long-eared Owls
i Asio otus) for a dachshund (Canis familiaris) . Therefore, the tendency to follow mam-
malian predators may exist independently of the cooperative feeding phenomenon among
Marsh Hawks and other avian predators. However, the tendency also could serve as the
behavioral basis for cooperative feeding between avian and mammalian predators when
the opportunity arises. — LeRoy W. Bandy and Barbara Bandy, Rt. 1, Box 75, Stetson,
Maine 04488. Accepted 4 March 1977.

Predation ecology of eoexistng Great Horned and Barn owls. — Food habits of
the Great Horned Owl (Bubo virginianus) and the Barn Owl (Tyto alba) are well studied
(e.g., Wdlson, Auk 55:187-197, 1938; Graber, Condor 64:473-487, 1962), but an emphasis
on feeding ecology and niche segregation is fairly recent (Marti, Condor 76:45-61, 1974).
This paper details some of the mechanisms facilitating coexistence of these owls during
the summer at Tule Lake National Wildlife Refuge, Siskiyou County, California.

Methods. — Observations extended from 17 June to 12 July 1975. Of the 107 knr study-
area, about half consisted of open water; the remainder included the eastern slope of a
large ridge where both owl species roosted on rock cliffs, a region of natural vegetation
along the base of the ridge, and agricultural fields to the east. The onset of owl activity
at 2 rock cliffs (northern and southern, 5.3 km apart) was recorded on alternate evenings.
Small rodents were trapped and tethered (with brass wire wrapped at the base of the tail)
on 2 dirt roads. 1 wdth telephone poles and 1 without, to test the importance of high
perches in the hunting patterns of the owls. Identity of predators was determined either
from direct observation with a night scope or observation of wing marks and footprints
around the kill. Kills of (juestionable identity were excluded. The presence of car and
observer did not constitute a new or unusual feature at either site, since parked farm
equipment is common along the roads. Habitat preferences and hunting patterns were
studied by driving through the area in a non-systematic pattern between 22:00 and 04:00
PDT. Twenty-six h of these observations were recorded over 17 nights.

Pellets were used to determine food habits and were collected at weekly intervals at
known owl roosts. Barn and Great Horned owl pellets were separated on the basis of size,
shape, firmness, and exact location of collection, as suggested by Moon (Trans. Kans.
Acad. Sci. 43:457-466, 1940); those of questionable origin were discarded.

March 1978 • GENERAL NOTES

135

r I i I I Ml I I I I rr I i i i i i i i i ii i

Barn Owl
first vocalization

Barn Owl
first flight

Great Horned Owl
first vocalization

Great Horned Owl
first flight

Li_J_l I I 1 I I 1 1 I I I I I I i I I I I I I I I

2000 2030 2100 2130 2200

Fig. 1. Activity onset in Great Horned Owls and Barn Owls roosting on 2 cliffs at Tule
Lake National Wildlife Refuge, measured by initial vocalization and initial flight from
the cliff. Vertical line = mean, horizontal bar 95% confidence limits of the mean,
top bar = northern cliff, bottom bar = southern cliff.

Statistical tests follow those described by Snedecor and Cochran (Statistical Methods,
Iowa State Univ. Press, Ames, 1967 ).

Food habits. — I analyzed 250 whole pellets and numerous pellet fragments containing
1003 prey items. Great Horned Owls averaged 3.83 and Barn Owls 2.42 prey items/pellet.
A significant difference existed between proportions of different prey taken by the 2
species ( x“ = 13.41, df r= 2, P <7 0.005 ) , although extensive overlap was evident (Table 1 ) .

Activity at roosting sites. — Four Great Horned Owls were resident at each of the 2
cliffs; Barn Owls numbered 25 at the southern cliff while 5 was the maximum heard at
ariy one time at the northern cliff. Initial vocalization and initial flight from the roost
were recorded as indicators of activity onset. Great Horned Owl activity onset, though
somewhat variable with respect to time (Fig. 1), was net significantly different at the 2
cliffs for initial vocalization (2-tailed t-test, t = 0.311, P>0.60) or initial flight (t =
0.338, P>0.60). Barn Owl activity, however, began significantly later at the northern
than at the southern cliff (Fig. 1) for initial vocalization ( t = 4.684, P < 0.001 ) and for
initial flight (t = 4.845, P< 0.001). Although data were limited. Barn Owls also ap-
peared to return to the roost earlier than Great Horned Owls over 4 mornings of observa-
tion. Generally they had left exposed perches for more protected roosts and their vocaliza-
tion level had dropped noticeably by the time Great Horned Owls arrived at the cliffs.

Roosting sites of individual owls were divided into 3 categories based on extent of ex-
posure. Barn Owls chose less exposed roosts significantly more often than Great Horned
Owls ( X' = 13.20, df = 2, P <0.005). Barn Owls typically roosted far back in protected
crevices or in deep holes where they were invisible from the road, while Great Horned
Owls perched on exposed rocks or ledges, or in large open holes.

VM)

I'HE WILSON BULLETIN • VoL 90, No. 1, March 1978

Table 1

Pfiey Items Identified in Owl Pellets Collected at Known Owl Roosts Within the

Study Area^

Great Horned Owl

Bam Owl

Number

Percent

Number

Percent

Prey species

of items

of total

of items

of total

Microtus

404

66.7

241

60.6

Perornyscus

167

27.6

147

36.9

Dipodomys

1.1

0.8

Sylvilagus

0.3

0.2

Euphagus cyanocephalus

0.3

Sorex

0.2

Tadarida brasiliensis

0.2

Mustela frenata

0.2

Unidentified bird

1.4

0.5

Unidentified small mammal

0.7

Insect

1.3

0.8

^ Includes 107 Great Homed Owl pellets, 143 Bam Owl pellets, and numerous pellet fragments
from both species.

Direct interspecific interactions were observed only twice; these consisted of single
Barn Owls harrying or diving at single Great Horned Owls at the southern cliff. Indirect
interactions occurred on at least 8 occasions when either the arrival of a Great Horned Owl
at one of the cliffs or the beginning of its vocalizations was accompanied by a decline or
brief cessation in Barn Owl activity and vocalizations. In addition, remains of at least 4
Barn Owls were found near the southern cliff under perches used by Great Horned Owls.
Thus not only competitive interactions, but also predator-prey interactions were occurring.

Hunting behavior. — Twelve kills of tethered prey were observed on the road having
telephone poles; of these, 8 were by Great Horned Owls and 4 by Barn Owls. All 6 kills
occurring on the road without poles were by Barn Owls. The difference between numbers
of kills by the 2 species at the 2 sites was significant ( corrected for continuity 4.640,
df = 1, P<0.05), with Great Horned Owls favoring the road having telephone poles.

Great Horned Owls made extensive use of telephone poles and to a lesser extent other
perches, while Barn Owls spent more time on lower perches, on the ground, or in flight
(Table 2). This necessarily limited the hunting hal)itat used by the larger species:
Great Horned Owls were never sighted in areas where perches w^ere not present. Barn
Owls showed a more uniform distribution throughout the area, although few were sighted
along the road at the base of the cliffs where the majority of Great Horned Owl sightings
were concentrated.

Search and attack behavior also varied between the species. Generally, Great Horned
Owls moved regularly and directly from one telephone pole to the next along a road,
spending from 1 to 59 min on a pole (X = 7.3, n m 38). When prey was sighted, a steep
downward flight was made, with the owl sometimes banking just before landing. Usually
the wings were flapped briefly on landing, after which no movement was seen for a period
of ’^/‘2 to 3 min until the owl took off again, flying directly up to one of the poles nearby.
Great Horned Owls were most often observed hunting alone, although groups of 2 or 3 owls

March 1978 • GENERAL NOTES

137

Table 2

Record of Barn Owls and Great Horned Owls Sighted During 26 ii
Through the Study Area Between 22:00 and 04:00^

Spent Driving

Great Homed Owl

Barn Owl

In flight
On perches:

Telephone

poles

Signposts

Ground

Other

Total

1 Observations before 22:00 or after 04:00 were excluded so that owls emerging from roosts
or returning in the morning would not bias data.

were twice seen moving from pole to pole together. In both cases vocalizations occurred
almost continuously between members of the group.

Barn Owls hunted primarily on the wing and occasionally from low perches. Hunting
flight was usually low, with a quick erratic wingbeat or, less frecjuently, a fast direct flap,
as described by Wilson (1938). This species was most often observed flying along irriga-
tion channels or over strips of natural vegetation on the levee paralleling the road. Sud-
den steep banking drops into the vegetation were common, and owls often emerged several
seconds later when unsuccessful.

Discussion. — Differences in hunting methods and habitat preferences result in reduced
spatial overlap, giving Barn Owls access to areas not normally used by Great Horned
Owls. These differences in hunting habits are probably physically based: the smaller

size and lighter wing loading of the Barn Owl may make hunting on the wing profitable
in spite of the energy expenditure, while the larger Great Horned Owl with its heavier
wing loading may be constrained to hunting primarily from perches ( Earhart and John-
son, Condor 72:251-264, 1970; Marti 1974).

Balancing this is the interactive dominance of the Great Horned Owl and its status
as a potential predator on the smaller owl. Inhibition of Barn Owl activity l)y Great
Horned Owl arrivals at the cliffs, a lack of Barn Owl sightings where Great Horned Owl
sightings were concentrated, and remains of Barn Owls found below Great Horned Owl
perches all point to the importance of this interaction. The selection of protected roost-
ing sites by Barn Owls is consistent with this, as is the delayed Barn Owl activity onset
at the northern cliff where Barn Owls were much less numerous and Great Horned Owl
activity was more prolonged and conspicuous.

Acknowledgments. — I wish to thank the U.S. Fish and Wildlife Service, 1). W. Ander-
son, N. K. Jacobsen, H. W. Li, D. G. Raveling, R. G. Schwab, and 1). S. Zezulak for
their assistance and encouragement. — Seri G. Rudolph, Dept, of U ildlife and Fisheries
Biology, Univ. of California, Davis 95616. Accepted 9 Sept. 1977.

THE WILSON BULLETIN • VoL 90, No. 1, March 1978

IBB

llosi ro<*or<ls f<>r the Siripod Ciiokoo from (>osta Rica. — The Striped Cuckoo
Mfapcra nacvia) is one of only 3 species of New World cuculids that exhibits obligatory
brood parasitism. Accounts of its reproductive behavior and host records have been pro-
vided by Ilaverscbmidt (.1. f. Ornitbol. 95:337-343, 1955; J. f. Ornitbol. 102:353-359,
1951) for Surinam and by Eriedmann (Ibis (13)3:532-538, 1933) for various parts of
South America.

Little is known about the habits of the species north of South America, although it is
common in many lowland areas of Middle America. Aside from a single instance of
parasitism on tlie Rufous-breasted Spinetail iSynallaxis erythrothorax) in Guatemala,
which was filmed by Hugh C. Land (Wetmore, Smith. Misc. Coll. 150(2) :132-135, 1958),
we can find no other host record for the species north of Panama. Wetmore (1958)
found “beautiful blue” eggshell fragments in the oviduct of a Striped Cuckoo he col-
lected at Buenavista, Chiricjuf, Panama on 1 March 1950, hut eggs of the Middle Amer-
ican race, T. n. excellens. otherwise appear to be undescribed.

While working near Rincon de Osa, Puntarenas Province, Costa Rica between Fehruary-
April, 1971, A. W. found the Striped Cuckoo to he common in second growth adjacent to
primary rain forest at elevations <20 m. On 10 April 1971 in a second growth area about
2 km W of Rincon de Osa, he found a nest of the Plain Wren i Thryothorus modestus)
which contained 1 wren egg and another egg believed to be that of the Striped Cuckoo.
The nest. ty})ical for the Plain Wren, was a retort-shaped structure composed of dry
stiff grass stems and grass heads and lined with a few feathers. It was placed 1 m off
the ground in the vertical fork of a small thorny shrub. The nest and eggs were collected
and are now in the collection of the Western Foundation of Vertebrate Zoology (WFVZ
no. .58440) .

Both eggs were fresh. The wren egg is white, subelliptical in shape (after Preston in
Palmer, Handbook of North American birds, vol. 1, Yale Univ. Press, 1952:13), some-
what glossy, and measures 18.75 X 14.79 mm. It is essentially identical to 10 other eggs
of this species from southwestern Costa Rica in the WFVZ collection. The presumed
Striped Cuckoo egg is medium blue (faded from bright greenish-blue when collected),
oval (Preston 1952), lacks gloss, and measures 23.43 X 16.46 mm. These details agree
with the description given by Hellebrekers (Zool. Med. Ryksm. Nat. Hist. Leiden, 24:
251-252. 1942) for 50 Tapera nacvia eggs from Surinam (range of measurements
18.7-23.5 X 14.1-17.3 mm).

Three other nests of Thryothorus modestus containing 2 eggs each were found in the
Rincon de Osa area between 21 March-19 April 1971, but none contained parasite eggs.
We are unaware of any previous record of Tapera naevia parasitism on Thryothorus
modestus. although Wetmore (1958) suggested that this wren might be a suitable host for
the parasite in Panama and described several instances of Striped Cuckoo parasitism on
a congener, Thryothorus rujalbus.

At a nearby southwestern Costa Rica locality, Sierpe, a small village near sea level, 13
km S of Palmar Sur, Puntarenas Province. L. F. K. found Striped Cuckoos to be unusually
common from March to November, 1970. The species was not reported from this area
prior to 1955 (Wolf, Condor 58:4€0-401), but it is apparently increasing in abundance
in soutlnvestern Costa Rica as primary habitats are reduced to second growth by agri-
cultural and lumbering activities.

On 31 May 1970 on tlie outskirts of Sierpe, L. F. K. dismantled 1 of the massive stick
nests of the Pale-hreasted Spinetail ( Synallaxis albescens) and found that it contained 2
spinetail eggs and another egg presumed to be laid by the Striped Cuckoo. The nest,
the only one of this species examined in the vicinity, was situated in a tangle of vines on

March 1978 • GENERAL NOTES

139

a fencepost standing in the mid(Jle of a dense Heliconia thicket. The incubating spinetail
was flushed from the nest.

The spinetail eggs were dull white when collected, hut acciuired a distinctly greenish
tinge after they were blown. They are short oval in shape < Preston 1962 » and have a
rough texture. They measure 19.40 X 16.01 and 18.92 X 15.90 mm. The probable
Striped Cuckoo egg is identical in color and shape to the 1 eollected by Williams at Rin-
con de Osa and measures 22.01 X 16.45 mm. The specimens are now in the W FVZ col-
lection ( no. 51515) .

While this is the first record for Synallaxis albescens as a host in Middle America,
it is known to he parasitized frequently by Striped Cuekoos in various parts of South
America (Friedmann 1933; Haverschmidt 1955).

The parasite eggs deserihed here are identified on the basis of circumstantial evidence,
hut we believe that their designation as Tapera naevia eggs is an accurate one. In color,
size, and texture they agree with j)uhlished deseriptions of the eggs of the 2 South
American races of the .Striped Cuckoo. Tapera n. naevia and T. n. chochi (e.g., Helle-
hrekers 1942; Friedmann 1933). Since there are only slight mensural and color differ-
ences between these subspecies and the Middle American T. n. excellens, it is rea-
sonable to expect that their respective eggs are very similar, at least in size. .Striped
Cuckoos were common at both Rincon de Osa and .Sierpe. and they w'ere occasionally
seen perched on the fencepost that supported the spinetail nest described from the latter
loeality. Finally, based on our joint experience with the nesting birds of Costa Rica and
an examination of the eggs of neotropical species in most major North American collec-
tions, we know of no other Middle Ameriean species that lays eggs of this descri{)tion.

Our fieldwork in Costa Rica was supported by the Western Foundation of Veitehrate
Zoology and Ed N. Harrison. — Lloyd F. Kiff and Andrew Williams. W estern Foundation
of Vertebrate Zoology, 1100 Glendon Avenue, Los Angeles, CA 90024. (Present address
AW : P.O. Box 23, Njoro, Kenya) . Accepted 10 Oct. 1977.

Ant-following: birds in South American subtropical forests. Ajiart from their
legendary aspects, army ants i Dorylinae) have attracted the attention of ecologists lie-
cause of the interactions between the ants and their associated bird followers (e.g., Willis,
Living Bird 5:187-231, 1966a; Oniki, Acta Amazonica 2:59-79, 1972). Hilty (Wilson Bull.
86:480-481, 1974) has called attention to the rarity of reports of birds associated with
army ants at higher elevations, and we report here on birds associated with 2 such ant
swarms.

On 16 and 19 April 1973, with R. Cochfeld and M. Kleinhaum, we visited Pichinde at
about 1700 m near the crest of the western Andes above Cali. Department of Valle.
Colombia. The vegetation and area have been described by Trapido and .'^an Martin
(Am. J. Trop. Med. Hyg. 20:631-641, 1971). On 19 April, on the steep slope in upper
subtropical forest, above a fast-rushing stream, we encountered a swarm of small black
army ants i Neivamyrmex sp. ). We remained with the swarm from about 08:30 to 10:00.
Our attention was attracted by the calls of Crested Ant-tanagers )Habia cristata) and
Lineated and Montane foliage-gleaners iSyndactyla subalaris and Anabacerthia striati-
collis) . We noted up to 10 of these ant-tanagers which foraged mainly between 1 and 2
m above the ground, moving along the edge of and in front of the swarm. They were
noisy, giving loud nasal jay-like calls as described by Willis (Condor 68:56-71, 19661) ) and
were quite animated, freijuently erecting or “flashing” their red crests, sometimes main-

THE WILSON BULLETIN • VuL 90, No. 1, March 1978

taining them erect for several seconds. The 2 si)eeies of foliage-gleaners were represented
at the swarm hy 2 individuals each. The Anahacerthia foraged in more or less upright
hranehes mainly above 2 m, while the 2 Syndacty/a remained closer to the ground and
called repeatedly. During our visits to Pichinde, we found the tanagers and Syndactyla
only at the swarm.

The following species were also persistent attendants at the swarm: 2 or 3 Black-billed
Thrushes (Tardus ignobilis) ; 1 female anthird (Myrmeciza spp., probably M. immacu-
lata), seen in low vegetation within 1 m of the ground; 1 male Slaty Antwren ( Myrmo-
therula schisticolor) seen rei)eatedly over the swarm; 1 male Plain Antvireo ( Dysitham-
nus mentalis) ; 1 Spotted Barbtail (Premnoplex hrunnescens) , foraging 1-3 m above
the ground; 2 (iray-breasted Woodwrens ( Henicorhina leucosticta) actively foraging
among tbe ants; 1 Rufous-naped Greenlet ( Hylophilus semibrunneus) .

Additionally, several species were seen only briefly near the swarm. For example, a
male Andean Cock-of-the-Rock (Rupicola peruviana) flew right into the area where the
Ant-Tanagers were calling, as if attracted to the commotion. We could not watch it in
the dense foliage and do not know whether it remained and fed. Several other species
appeared to be associating with the bird flock, rather than benefiting from the ants.
Willis (1966b op. cit.) and Oniki (Condor 73:372-374, 1971) noted that wandering bird
flocks often join ant-following flocks without actually using the ants. Thus the coalescence
and disintegration of flocks of different social structure, feeding in different manners
and moving at different rates, may occur regularly in subtropical and tropical forests.

Hilty (op. cit.) reported 2 species of Tangara tanagers attending mid-elevation (900-
1500 m) swarms in Colombia. At Pichinde we recorded 8 species of Tangara within 300
m of the swarm (some much closer), but none actually associated with it. Significantly,
the only bird common to our flock and Hilty’s was the Myrmeciza, which Willis (pers.
comm.) has found to be a regular ant-follower elsewhere. Willis (1956b op. cit.) gave
a detailed account of tbe ecology and behavior of the Crested Ant-tanager, and the hab-
itat at Pichinde seems to be characteristic. None of the 16 mixed flocks in which Willis
saw H. cristata were attending ants, and S. Hilty and R. Ridgeley have seen it at
Pichinde in the absence of ants. Although this appears to be the first report of H.
cristata following ants, some other members of the genus do so regularly.

Of the 45 species which Willis listed in 16 flocks, 10 occurred in the flock we observed.
Of these 10, the wood wren and Slaty Antwren occurred in 9 and 8 of the 16 non-ant-fol-
lowing flocks studied by Willis, raising the question of whether there is a consistency to
mid-elevation flocks quite apart from a common attraction to ant swarms. The Golden
Tanager {Tangara arthus), one of the commonest forest birds at Pichinde, was not seen near
the flock, and Willis found it in only 1 of 16 flocks. Hilty (op. cit.) indicated that the
small black ants he encountered in the upper Anchicaya Valley w'ere “presumably L.
l=Labidus] praedator.” Tbe ants from Pichinde were identified as Neivamyrmex sp. by
H. Topoff. Although there have been no previous reports of birds following ants of this
genus, it is likely that Neivamyrmex swarms will be attended when appropriate bird
species encounter them.

On 30 January 1974, at 1550 m altitude in subtropical forest above the headquarters
of Par(iue Nacional de Rancho Grande, Aragua, Venezuela. R. Gochfeld, M. Kleinbaum
and M. G. found an ant swarm attended by 4 Black-faced Antthrushes (Formicarius
analis) , and 1 Short-tailed Antthrush ( Chamaeza campanisoma) . A Strong-billed Wood-
creeper ( Xiphocolaptes promeropirhynchus) spent about 5 min foraging on vertical
trunks 2-10 m above the swarm. Unlike Pichinde, no other species appeared near the

March 1978 • GENERAL NOTES

141

swarm, and the 3 species attending the swarm were silent for the entire time. The ant
species was not identified.

In conjunction with Hilty's observations lop. cit.) it appears that the paucity of re-
ports of ant-attending birds at high altitudes may reflect the relative scarcity there of
Doiyline ants, and that where such ants occur, one may anticipate that some bird species
will attend them. It is unlikely that professional ant-followers (in the sense of \^’illis,
Ecology 47:667-672, 1966c; Oniki and Willis. Acta Amazonica 2:127-151, 1972), could
maintain themselves more than marginally at such altitudes. Willis I pers. comm. ) notes
that Myrmeciza immaculata is probably such a professional, but its altitudinal range is
mainly below 1700 m. In view of the scarcity of raiding ants, ant-attending birds of sub-
tropical forests should be mainly non-professional opportunistic species, offering an in-
teresting chance to study their behavioral interactions in the absence of professionals.
Veiy often the most interesting insights into complex ecologic situations come from ob-
sening phenomena at the extremes of a range where atypical events are likely to occur.
Further investigation of birds at high altitude swarms will provide opportunities to
extend observations made at low elevations.

Our field work in Colombia benefited in many ways from the kind assistance of the
late Dr. F. Carlos Lehmann whose death has meant a severe loss to neotropical ornithology.
Field work at Rancho Grande was made possible by Dr. Gonzalo Medina, and we very
much appreciate the adnce and assistance of Paul Schwartz. Robert Gochfeld and
Michael Kleinbaum participated in both trips. We thank Steven Hilty, Edwin 0. \^’illis,
and Yoshika Oniki for comments on the manuscript. — Michael Gochfeld, Field Re-
search Center, Rockefeller Unit., Millbrook, XY 12545, and Gly Tudor, 380 Riverside
Drive, A Y 10025. Accepted 17 Dec. 1976.

Fishing behavior of Black and Turkey vultures. — Black and Turkey vultures
(Coragyps atratus and Cathartes aura) are usually characterized as carrion feeders, though
both species have occasionally been observed taking live prey ( e.g.. Bent. L .S. Natl. Mus.
Bull. 167, 1937; Mueller and Berger. Auk 84:430. 1967; Gladding and Gladding. Condor
72:24-4-245, 1970; Bang, J. Morph. 115:153-164. 1968 1. hile both species are known
to include fish in their diet, it has usually been assumed that the fish were obtained as
carrion. Bendire ( U.S. Natl. Mus. Spec. Bull. 1. 1892 1 however collected a Turkey Vul-
ture which had a crop gorged with "fresh" small minnows, at least suggesting the pos-
sibility that the fish were taken alive. hile discussing these species, we discovered that
each of us had observed vultures fishing and decided to collaborate in the presentation
of our observations. Our observations include apparent fishing by Black Vultures in
Virginia and Mississippi and by a Turkey Vulture in Florida.

\ irginia. — Just before dusk on 2 January 1975, in the Radford Army Ammunition
Plant, 14 km west of Blacksburg. Virginia. Prather and Conner observed Black Vultures
engaged in a behavior which looked like fishing. Beneath a large roost ( Prather et al.,
ilson Bull. 88:667-668. 1976) on the banks of the New River, 5 vultures stood along
the bank and peered into the water. Three others on fallen limbs jutting out over the
river assumed positions much like that of the Green Heron ( Butorides firescens; Bent.
L.S. Natl. Mus. Bull. 191, 1950 » when fishing from a perch. A continuous rain of ex-
crement from the roosting birds above peppered the river. This may have attracted fish.
From a distance of 0.5 m above the water, one of the Black Vultures suddenly pushed
with its legs and dove into the river. The bird’s head and body were completely sub-

THE WILSON BULLETIN • Vol. 90, No. 1, March 1978

1 12

Fig. 1. Black Vultures “fishiufi” at a spillway, Bluff Lake, Noxubee National Wild-
life Refuge, Mississippi.

merged, though it held its wings out of the water. The vulture immediately surfaced and
made its way to shore. The success of the vulture's efforts could not he determined.

This behavior, minus the culminating immersion, was again witnessed on 4 suhsecjuent
occasions in the month that followed. The river hank beneath the roost was searched for
fish remains or other sign, hut none were found. However, all perches which jutted low
(.5 to 1.0 m) over the water and were greater than 8 cm in diameter were worn smooth
on the top surface, presumably by perching vultures.

Mississippi. — At 09:00, on 11 May 1976. Jackson observed 19 Black Vultures standing
in the watcu or on concrete next to the water at the spillway of Bluff Lake on Noxubee
National ildlife Refuge. Oktibbeha County. Mississippi. One of the vultures in the
water was feeding on the head of a large dead fish that had apparently washed over the
spillway. Others stood motionless facing upstream or across the current (Fig. 1). Oc-
casionally one would grab at something w4th its beak and one foot. One vulture captured
a 6-8 cm live fish in this manner. Humans frecpiently fish near the spillway and un-
desirable fish — often injured or dead — are tossed out on the hank or hack into the water.
Vultures previously had been seen feeding on dead fish on the hank. Injured or dead fish
in the water would he a little more difficult for the vultures to t)htain hut might lie
relati\ely easy to catch as they washed over the spillway.

March 1978 • GENERAL NOTES

143

Florida. — At about 17:20, on 20 May 1976. in the East ilderness area of Fish-eating
Creek Campground, Glades County, Florida. Gahy watched a Turkey Vulture from about
10 m as it landed on the sloping hank of Fish-eating Creek. The vulture walked down
the hank to the creek and into the water to a depth at which the water was almost in
contact with its belly feathers. Then the bird began, apparently, to search for something
in the water. It made several stabs at the surface with its hill and. at the same time,
spread its wings as if for balance. T^'hen it raised its head it had a wiggling fish, ap-
proximately 10 cm long, in its hill. The vulture walked hack to shore where it consumed
its catch. After eating the fish, the vulture reentered the water and made additional at-
tempts at “fishing" which were not successful.

These instances, while probably not representative of typical vulture behavior, indicate
a potential for using different foraging tactics. At a time when “traditional" \ulture food
may he becoming less available, perhaps more aggressive and more opportunistic foraging
tactics will he selected for.

\^'e wish to thank Oscar Owre and Ren Lohoefener for helpful comments on various
parts of this note. — Jerome A. Jackson. Department of Zoology, Mississippi State Uni-
versity, Mississippi State 39762: Irvine D. Prather and Richard N. Conner, Depart-
ment of Biology, Virginia Polytechnic Institute and State University, Blacksburg 24061:
AND Sheila Parness Gaby. Department of Biology, University of Miami, Coral Gables, FL
33124. Accepted 18 Dec. 1976.

new hybrid warbler combination. — An unusual warbler captured in a mist net
on 12 October 1967 at Nantucket. Nantucket Co., Massachusetts, was preserved as a speci-
men by Baird who suspected that it was a hybrid. It was a female with an incompletely
ossified skull and weighed 12.0 g. On comparison with other preserved material, the bird
was tentatively identified as a hybrid \ ellow-rumped ( Myrtle >. Dendroica coronata. X
Bay-breasted, D. castanea, Warbler. It was similarly and independently identified by
Banks after comparison with material in the National Museum of Natural History. This
hybrid combination was not mentioned by Gray (Bird Hybrids. Commonwealth Agric.
Bur., Farnham Royal. Bucks, England, 1958) and has not. to our knowledge, been
reported in subsequent literature. The following comparative description is based on
immature (first fall) females of the presumed parental species.

The hybrid Myrtle X Bay-hreasted \^’arhler (USNM 567882) is ver> similar dorsally
to a Blackpoll \^'arhler ( D. striata) and might easily he mistaken for that species at a
glance. It is. however, slightly darker and somewhat grayer. The hack and nape color
is intermediate between the rather bright yellowish-green of the Bay-hreast and the
brownish of the Myrtle. The crown of the hybrid is lighter than the hack, approaching
the color of the Bay-breast. Feathers of the crown have, distal to the basal gray area,
a small spot of white along the rachis and a suffusion of yellow extending onto the vanes;
the tips of these feathers are green. Neither the white nor the yellow is as extensive as
in the Myrtle arhler. There is a yellowish cast on the rump feathers of the hybrid,
but none of these feathers has the bright yellow tip characteristic of the Myrtle. The
upper tail coverts are edged with silvery gray, as are those of the Myrtle, in contrast to
the green-tipped gray coverts of the Bay-hreast.

The ventral body surface is essentially plain, with a faint hand of dark spots across
the breast. Some of the flank feathers, particularly the more posterior ones, have dark

141

THE WILSON BULLETIN • Vol. 90, No. 1, March 1978

shaft streaks. In both of these characters, the hybrid is similar to the Myrtle, although
not as extensively marked. The abdomen is white as in the Myrtle. Otherwise the ventral
coloration (including the under tail coverts) is huffy as in the Bay-breast, although
paler. This huffiness rules out I), striata as a possible parent. There is a very slight
yellowish cast on some mid-flank feathers but no indication of yellow on the throat.
There are large white spots on the inner vanes of the outer 2 rectrices on either side,
and a smaller white spot on the third rectrix on the right. The Bay-breasted Warbler
typically has such spots on the outer 2 rectrices, the Myrtle on the outer 2 or 3. The
small amount of white on the rectrices and the lack of yellow on the throat, as well as
consideration of geographic ranges, eliminates the Audubon type of Yellow-rumped
Warbler from consideration as a possible parent.

Measurements < by Banks) of a small series of each parental form indicate that although
there is considerable overlap, the Myrtle Warbler averages slightly smaller than the Bay-
breast in wing and tarsus length and the Bay-breast is slightly smaller in length of the
tail and middle toe. In each of these measurements, the hybrid is smaller than the mean
of the smaller species, although within the range of variation of the smaller or of both
parental forms. There is little overlap in the length of the bill, measured from the anterior
edge of the nostril, in the parental species, the Bay-breast being longer billed. The
presumed hybrid is very near the mean of the Myrtle Warbler in this measurement. Thus
the hybrid is in most respects smaller than either parental species, in contrast to the
intermediate size often noted in hybrids.

Parkes (Condor 63:348-449, 1961) has pointed out that all known wood warbler hybrids
are either intergeneric or between members of closely related species-pairs. This seems
to be the first report of an intrageneric hybrid between species not members of a species-
pair.

In characterizing the genus Dendroica, Ridgway (U.S. Natl. Mus. Bull. 50, pt. 2, 1902)
noted that the wing is rather pointed with the “four outermost primaries abruptly
longest . . .” except in the distinctive rounded-winged Antillean species and in D. mag-
nolia, in which the outer primary is reduced. He did not mention that in 3 species—
D. castanea, I), striata, and D. caerulea — only the 3 outermost primaries are abruptly
longer than the inner ones ( Z). fusca nearly approaches this condition). The hybrid is
similar to D. coronata in having a wing tip made up of 4 long feathers rather than 3
as in D. castanea. — Richard C. Banks, Div. of Cooperative Research, U.S. Fish and
U ildlife Service, Washington, D.C. 20240, and James Baird, Massachusetts Audubon
Society, Lincoln, MA 01773. Accepted 26 Sept. 1976.

ORNITHOLOGICAL LITERATURE

Sexual Size Dimorphism in Hawks and Omls of North America. By Noel F. R.
Snyder and James W. 'Vi’iley. Ornithological Monographs No. 20, 1976 :vi + 95 pp. Amer-
ican Ornithologists' Union. $7.50. — The chief original contributions of this paper are (1)
a table summarizing food habits of all species of hawks and owls in North America for
which significant data are available in Fish and Wildlife Service files or published litera-
ture; i2) a table of dimorphism indices and the mean measurements of wing chord,
culmen, and weight on which these indices are based; (3) extensive data on food,
feeding rates and timing of reproductive losses at nests of the Sharp-shinned Hawk ( both
Puerto Rican and mainland subspecies). Cooper’s Hawk, Goshawk, and Red-shouldered
Hawk; (4) excellent new data demonstrating substantial sexual differences in diet in
Sharp-shinned and Cooper’s hawks; and (5) 4 bivariate scatter diagrams that in-

dividually relate average dimorphism in both hawks and owls to % birds (rs = .79), %
mammals <rs = .06), % lower vertebrates irs = .16), and % vertebrates (ts = .43) in
the diet. The correlation of dimorphism with % birds is highly significant; that with
% vertebrates is weakly significant, resulting mainly from the inclusion of the component
of avian prey. From this the authors conclude (p. 9) that. "The correlation between size
dimorphism and taking of avian prey is sufficiently strong that we consider it to be the
most important fact to be handled by any hypothesis concerning the function of size
dimorphism in raptorial birds.”

On the basis of these findings, Snyder and Wiley present a novel refinement of existing
theoiA' that relates sexual size dimorphism to adaptively-broadened intraspecific niches.
Specifically, they propose that substantial dimorphism in bird-feeding raptors has been
selected for because it reduces intra-pair competition and allows a wider range of food
sizes to be taken during the crucial period late in the breeding cycle when both sexes
are foraging and when such predators are food stressed because of reduced bird popula-
tions. Young of these birds of prey, requiring extended practice to develop the expertise
necessar\- to capture birds, may be especially vulnerable to food shortage during their pro-
longed post-fledging dependency period.

Although the authors achieve only moderate success in supporting their argument with
field data, especially those obtained from nesting accipiters, even when their efforts
lead to rather equivocal results one gains the impression that they are seeking answers
in the correct places. It is unfortunate that the general paucity of published information
on feeding ecology of hawks and owls late in the breeding season renders their hypothe-
sis difficult to test with information from other species. Nonetheless, Snyder and iley’s
stimulating thoughts on dimorphism deserve continuing consideration as new information
accumulates.

In addition to a thorough review of ideas proposed in the past for the sexual size dif-
ference in raptors, they also discuss the related (juestion of why the female is usually
the larger sex. After briefly entertaining the notion that reversed dimorphism "is pos-
sibly a chance effect,” they conclude, in at least partial agreement with Amadon ( Raptor
Research 9:1-11, 1975), that the reversed nature of the dimorphism is likely to he “re-
lated to advantages in copulation, incubation, brooding, and nest defense for large
females.”

The most impressive aspect of this monograph is the thoroughness and balance of the
discussion, which interweaves both old and new explanations for dimorphism in a most
satisfying way. The many alternative views already in the literature are weighed fairly,
with no apparent urge to seduce the reader into following poorly-illuminated paths. In

145

14()

rilE WILSON BULLETIN • VoL 90, No. 1, March 1978

Table

Kki.ationshii’ Bktwkkn Sizk Dimorphism and

Hawks and

% Birds i.n Diet in North American
Owls'

Taxa

included

Type of Number of

dimoriihism index siiecies

Statistical

significance

Hawks and owls

Average"

0.79

P < 0.001

Owls

Average

0.71

P -- 0.001

Hawks

Average

0.86

P < 0.001

Hawks below I). I. of 7.5

Average

0.13

P > 0.100

Hawks above I). I. of 7.5

Average

0.71

P = 0.010

Hawks and Owls

Body weighU

0.93

P < 0.001

Owls

Body weight

0.54

0.01 < P < 0.025

Hawks

Body weight

0.81

P < 0.001

Hawks below D. I. of 8.1

Body weight

0.09

P > 0.100

Hawks above D. I. of 8.1

Body weight

0.36

P > 0.100

1 Based on analysis, with the Spearman Rank Correlation Test, of data from Table 1 of Snyder
and Wiley.

“ “Average Dimorphism Index” of Snyder and Wiley is computed as the mean of separate in-
dices calculated for wing chord, cidmen, and body weight.

® Based on cube roots of mean weights of males and females.

addition to discussion of the possible dimorphism-food relationship, the paper is replete
with ad hoc hypotheses involving ether ecologic features and their possible relevance to
the size difference between the sexes. Thus, correlations are attempted between de-
gree of size dimorphism and incidence of double-brooding, degree of coloniality. polyg-
yny, and sex ratio in various raptorial species, including those that do not fit neatly
into the apparent relationship of degree of dimorphism and percent avian prey. To be
sure, many of these comparisons do not yield convincing correlations, suggesting that the
explanation underlying the evolution of dimorphism could be more complicated than
researchers in this area have admitted. If the selective pressures that influence dimor-
phism vary among species, through time, and by geographic region, as could well be the
case, then the data necessary to expose the real biologic correlates of sexual size dif-
ference will need to be correspondingly refined and extensive. Those who interpret this
conclusion as an unnecessary retreat into complexity are reminded that it also could
represent an advance into reality. Simple answers are unlikely to questions that have
puzzled naturalists for over a century.

Because the authors do not separately analyze the various groups of raptors included
in their biologically important and statistically significant plotting of % birds in diet
against average dimorphism ( Fig. 1, p. 7 ) , I have taken the liberty to examine their data
in this light. The results of my comparisons are presented in Table 1. Of immediate in-
terest is that owls follow the same trend as hawks even though they do not reach the
striking indices of dimorphism shown by the latter group. Inspection of Snyder and
Wiley’s plotting of the hawks alone exposes yet another area of concern. Below an aver-
age dimorphism index of approximately 7.5. only species that feed to a minor extent on
birds are represented and no correlation of sexual size difference with the taking of avian
prey is seen (rs:=0.13). Above an average dimorphism index of 7.5. however, strong
correlation is demonstrated. Thus, a gradual increase of dimorphism with increasing avian

March 1978 • ORNITHOLOGICAL LITERATURE

147

feeding is not seen in approximately onedialf the species of hawks. But when dimor-
phism reaches a certain (threshold?) level, the sexual size difference is dramatically
correlated with percent birds in the diet. Most impressive of all is that the highest levels

of dimorphism in owls occur in the same region of the scale where the hawks shift from

non-correlation to correlation with avian diet. Comparison of these trends in dimorphism
with those of birds of prey of other geographic regions should he most rewarding. Space
does not permit speculation here on the reasons for these patterns in North American
species.

In the opinion of many, body weight is far superior to other rrreasurernents as an as-
sessment of overall size. Thus, I was disappointed to note that Snyder and Wiley base
all their size comparisons on an average dimorphism index calculated as the mean of 3
separate indices, those of wing chord, culmen, and weight. For many species the indices
are rather similar and nothing would appear to he gained by using the body weight in-
dex. But for others the difference is dramatic and an index obtained by averaging simply

conceals the true magnitude of the size difference between the sexes. For example, in

the Great Gray Owl the index based on weight is 10.9. versus 6.3 for the average index.
Comparable figures for the Shoit-eared Owl are 6.1 versus 2.5. Therefore, comparisons
of dimorphism indices based on body weight with % avian food are of interest, and I
present these data in the bottom half of Table 1. In general, the findings compare favor-
ably with those derived from the average index. The combined groups of birds of prey
show a substantially higher correlation than when the average index was used. In-
dependent comparisons of owls and of hawks again provide statistically significant
Spearman correlation coefficients. Once more a break is evident in a plotting of the
hawks, in the vicinity of dimorphism index 8 and between 10 and 20% birds on the
scale. Surprisingly, when the 2 groups on each side of the shift are analyzed separately,
neither correlation is statistically significant. Although the reasons for this are not clear,
I suggest that in hawks, the dimorphism index based on weight correlates well with %
of birds taken as food, mainly because the means of 2 rather independent groups are
correlated. ithin either group, however, the scatter is substantial and no significant cor-
relations are evident. Thus, there are still many puzzling aspects of the relationship be-
tween degree of dimorphism and diet in addition to those illuminated by the authors.

In conclusion, my overall impression of this monograph is ver> favorable. Snyder and

iley are to be commended for enlivening the rather static theoretical framework upon
which discussions of sexual dimorphism have rested in recent years. \^’hether arguments
supported principally by data from accipiters will deserve generality depends upon in-
formation yet to be gathered from other species already studied superficially as well as
from additional birds of prey inhabiting geographic regions beyond North America.
Hopefully many of their suggestions and tangential leads will be followed by the current
army of raptor enthusiasts now alerted to this fertile ground. — Ned K. Johnson.

Ornithological Gazetteer of Paraguay. By Raymond A. Paynter. Jr., and Alastair
M. G. Caperton. Museum of Comparative Zoology, Cambridge, Mass., and Field Museum
of Natural History, Chicago, 111., 1977: iv -f -J3 pp.. 2 maps, paperbound. SL75. Order
from Bird Dept.. Mus. Comp. Zook. Harvard Univ.. Cambridge, Mass. 02138 or Bird
Division. Field Mus. Nat. Hist.. Chicago, 111. 60605. — This is the third, and much the
smallest, of the useful series of gazetteers of Neotropical localities where birds have been
collected or observed. Melvin A. Traylor, Jr., the coauthor of the volumes on Bolivia

148

rilE WILSON BULLETIN • VoL 90, No. 1, March 1978

and Ecuador, continues (with Paynter) to edit the series, but the present volume’s
authorship changes to Paynter and Caperton; the latter is credited in the editors’ intro-
duction with having prepared the preliminary draft. The general format remains the
same as in the earlier volumes. The outline map (printed on the back cover as well as
on p. 43) reflects recent changes in the boundaries of the departments of Paraguay, which
in some instances are (juite different from those shown in standard atlases such as that
published by The Times of London. The map on p. 36 pinpointing collecting localities
shows the astoundingly small area of Paraguay from which birds have been obtained.
Most of the localities lie along the major rivers, except for a belt across the Chaco and
a scattering in the north and central portions of the area between the rfos Paraguay and
Parana, in the eastern half of the country. The northern half of the Gran Chaco, perhaps
one-third of Paraguay, remains ornithologically unexplored; it is hardly surprising that
most of the collecting localities of the newly discovered third species of living peccary,
Catagoniis wagleri (Wetzel, Bull. Carnegie Mus. Nat. Hist. 3, 1977), lie within this little-
known region.

As in the case of the gazetteer for Ecuador, the authors have attempted a complete
bibliography. At least one major paper was omitted: “Catalogo sistematico de las aves del
Paraguay,” by B. Podtiaguin (Rev. Soc. Cient. del Paraguay 5 (5), 1941), which in-
cludes a detailed review of Cuculidae and the description of a new subspecies, Crotoph-
aga ani lapchinskyi (p. 90) from Villa Hayes: Rio Verde, kl. 4, Chaco. In the first
20 pages of this paper are listed at least 9 localities that are not in Paynter and Caper-
ton’s gazetteer (Bernal-cue, Camacho, Colonia Elisa, Colonia Santa Lazara, Estancia
Suhin, Fortin Page, Puerto Juan Barbero, Rio Confuso, and San Ignacio).

As I stated in my review of the Ecuador gazetteer (Wilson Bull. 89:638-639, 1977), I see
no reason for haste in preparing these publications, which should be as nearly definitive
as possible when issued. The editors continue to deny themselves the assistance available
from colleagues working on Neotropical ornithology, by their failure to circulate the
manuscripts of their gazetteers in advance of publication. — Kenneth C. Parkes.

The Wrens. Record # ARA-2. Produced and narrated by John William Hardy.
Principal recordist, Ben B. Coffey, Jr. Produced in the Bioacoustic Laboratory of the
Florida State Museum. Published privately by John William and Carol K. Hardy,
Gainesville, Florida, 1977. One LP record in jacket. |6.00. — This is Hardy’s second
record, and it is a notable success. As with his first record (reviewed by me in Wilson
Bulletin 88:525-526) his objective is not only to provide listening pleasure but also to
educate the listener. In this one there is rather more emphasis on pleasure and less on
education, but the same format is used. Written information is confined to the back of
the jacket, while a considerable amount of spoken information accompanies each record-
ing. Forty-three of the 60 wren species are presented — a worthy achievement in itself —
and in most cases there are several cuts for each species, often from different localities.
These are listed on the jacket, together with the bare details of the recording. The rest
of the jacket contains a general introduction to the family, a discussion of wren vocal-
izations, brief taxonomic comments, and a list of species not on the record. In the
spoken commentaries there is no standard species treatment (nor need there he) ; Hardy
merely highlights the most interesting points about each recording. Range and habitat,
and sometimes plumage notes, are given for each species, but most of the comments con-
cern the vocalizations. Topics discussed include song type (solo/duet), function.

March 1978 • ORNITHOLOGICAL LITERATURE

149

geographic variation, and taxonomic relationships, with a frequent nod to aesthetics.
For Hardy the wrens are “perhaps the ultimate songbird family,” and the record goes
far toward proving him right. He believes that wren voices “may be the evolutionary
counterpart in voice to the birds of paradise or the hummingbirds.” It is clear that he
is completely eharmed by wren voices — but so is everyone who has birded in the neo-
tropics, myself included. None who listen to this record can doubt that the wrens are
among the world’s greatest songsters. Hardy’s original objective was to display some
of the “stunners” in the family, but as an ornithologist he was also concerned to “sur-
vey as mueh of the family as possible, to give scientists food for thought about systematic
relationships and ecological aspects of bird voice.” He succeeds in all these objectives.

Most of the recordings are of the highest quality. Some less good ones are included,
either to illustrate some point or because they are the only recordings of a particular
species, and this is as it should be since the author is trying to inform as well as enter-
tain. This record marks the emergence of Ben B. Coffey, Jr. as a top-notch field record-
ist. Two-thirds of the recordings are his, and they are uniformly superb. He travelled
far and wide to obtain them, and is to be congratulated on a fine achievement. This is
not to imply that the other recordists have not also made noteworthy contributions —
credit is due to John Arvin, Luis Baptista. Richard Bradley, John Fitzpatrick. Michael
Gochfeld. William Gunn, Charles Hartshorne. P. P. Kellogg. David Lee, Ted Parker,
and Paul Schwartz, while Hardy himself contributed one of the recordings.

The jacket cover features a simple but charming black-and-white drawing by Richard
Bradley of 2 wrens, together with sonagrams of their voices.

No production is without defects, and the principal one here concerns the spoken
commentary, which is much too long. The same problem marred Hardy’s first record,
where there was almost more commentary than bird song. Here he has cut back the
human voice somewhat, but not nearly enough. It is not without reason that nearly all
bird records present the commentary in written form, either on the jacket or in an ac-
companying booklet. The human voice palls on repetition while bird songs do not; the
space taken up by the human voice can be used for more bird songs; and information
retrieval from a spoken commentary is extremely difficult. To find out what Hardy says
about a given species you either have to play through the whole side ithe mateiial is
not divided into bands as in many records) or take a stab (literally) at finding the right
spot by dropping the needle on the record. Your record will likely soon be covered
with scratches. Hardy is aware of these objections, yet defiantly presses his maverick
approach. Hear him you will, whether you like it or net. He suggests that anyone who
gets tired of the commentary can make a tape copy of just the birds, to which I can
only reply, “Thanks a lot ! ”

Hardy says it is possible that Coffey’s recording of Cistothorus platensis from Colom-
bia may in fact be C. apolinari, whose voice was not known at the time, but I can con-
firm that it is in fact platensis. A similar recording made by Michael Gochfeld near
Bogota turned out to be C. platensis tamae, whereas C. apolinari, which we both recorded
elsewhere, has a totally different voice.

The jacket is not without faults. Sloppy typesetting in the genera Odontorchilus and
Hylorchilus has resulted in duplication and confusion. Timberline and Mountain wrens
are nos. 34 and 35 on the jacket, but on the record the Mountain \^’ren comes first.
And where are the cuiious Colombian localities “Above and W. Uribe” listed for
Thryothorus genibarbis'l I found Uribe in my atlas; perhaps W. Uribe is a western
suburb. I could not find Above (Above? ) ; but then, some Colombian villages are
very small.

150

THE WILSON BULLETIN • VoL 90, No. 1, March 1978

In spite of some drawbacks this is an important record. A lot of voices are here pul)-
lislied for the first time; Hardy provides much good information; and above all, the
songs of the wrens are sheer delight. Eor anyone interested in bird voices, this record
is a “must.” — STifAKT Kkitii.

Vkktkhkates of Elokida. By Henry \I. Stevenson. University Presses of Florida, Gaines-
ville. 1976:607 pp.. 11 plates, 15 figures, 68 maps. $35.00. — The subtitle, identification
and distribution, points out the author's objectives, namely to provide in one hook the
means to identify specimens of all the land and freshwater vertebrates known to occur
in Florida, and to outline their distributions. The hook is organized into 4 chapters;
an introduction, the keys, the species accounts, and descriptions of technicjues for col-
lecting and preserving vertebrates. A glossary, which is subdivided into sections on
fishes, herptiles, birds, and mammals, and an 18 page bibliography, which includes
references to the first occurrences in Florida of many vertebrates, are useful. The illus-
trations, though functional, are uniformly unesthetic. In all, 880 species are included.
The author has done well at keeping the contents updated. He treats numerous exotics
that have established feral populations recently, and the many accidentals that have been
recorded in recent years.

In addition to students, teachers, and amateur naturalists, the ever-growing group of
environmentalists will find the book useful. Unfortunately its price will prove too high
for many. I have tested the keys with a few specimens from all 7 classes. Based on this
admittedly insufficient sample, the results are good. To reach the proper species, one
must first correctly establish the class, order, and family. I like this re(}uirement because
it reinforces knowledge of vertebrate classification. Keying birds requires total length
measurements. In the introduction the author defends using this approach, although
he has failed to convince several colleagues, including me. — Glen E. Wo(jlfenden.

Wildfowl of Europe. By Myrfyn Owen, illus. hy Hilary Burn and Joe Blossom. Mac-
millan London Ltd., 1977:256 pp., 55 color plates, many line drawings and maps. £.12.00.
— This is an authoritative but nontechnical review of the ducks, geese, and swans of Eu-
rope. The author. Conservation Research Officer for the Wildfowl Trust at Slimbridge,
England, has combined a concise but smoothly written text with an unusually fine col-
lection of illustrations to produce a work that should please both the reader's scientific
and esthetic appetites. The first quarter of the book is devoted to 7 chapters dealing
with general aspects of waterfowl biology, including relationships, environment, popula-
tion structure, banding studies and migration, behavior, and relations with bumanity.
The hulk of the work, how'ever, is concerned with individual accounts cf the 55 European
species. Each account covers identification, voice, breeding behavior, distribution, feed-
ing babits. and other subjects of interest. Appendices summarize information on nests,
eggs, and young; winter weights and measurements; winter foods; and breeding records
and special recjuirements of birds in captivity. Much of this information is from previ-
ously un{)ublished records.

There are 55 full page color plates painted by Hilary Burn. These illustrate the various
species in natural jioses, showing sex, age, and seasonal plumage differences, as well as
downy young. These paintings are extremely lifelike and attractive. There are also many

March 1978 • ORNITHOLOGICAL LITERATURE

151

skillful line drawings by Joe Blossom, and a full set of range maps. The hook includes
a foreword hy Peter Scott and a bibliography.

Wildfowl of Europe is a handsome addition to the extensive illustrated literature of
this most popular group of birds. — Robert J. Raikow.

Ontogeny and Phylogeny. By Stephen Jay Gould. The Belknap Press of Harvard
University Press, Cambridge, Mass., 1977 : 501 pp. $18.50. — Generations of students
have learned that “ontogeny recapitulates phylogeny.” The tenacity of this concept arises
not only from the axiomatic finality of the statement, hut also because there is clearly
sonic kind of relationship between the stages of embryonic development and the patterns
of similarity between organisms. In this hook Stephen Jay Gould reexamines the rela-
tionships between ontogeny and phylogeny and attempts to provide a new assessment
of their significance. For the most part he succeeds admirably, though the complexity
of the result precludes any simplistic summary slogan comparable to Haeckel's pithy
aphorism. The first part of the hook reviews the history of ideas about ontogeny and
phylogeny, while in the second a somewhat eclectic, but basically new theory is de-
veloped.

From Aristotle to tlie Naturphilosophen of the late 18th and early 19th centuries
various analogies were drawn between the stages of ontogeny and some scale of organic
perfection. Aristotle saw' in human development the successive levels of nutritive (cor-
responding to plants), sensitive (animal) and rational (human) beings. Later philoso-
phers developed detailed scales of being from lower to higher forms. Although they
form a background to the recapitulation theory, these concepts were mostly idealistic
rather than evolutionary. The great 19th century embryologist Karl Ernst von Baer is
often regarded as a supporter of recapitulation, an idea encouraged hy Haeckel, hut
(iould makes clear than von Baer actually opposed Haeckel, and argued that development
is a process of differentiation rather than recapitulation.

With the advent of evolutionary theory, recapitulationists developed a model in which
animals evolved hy adding new stages to the ends of unchanged ancestral ontogenies. At
the same time these ontogenies became condensed in duration so that successive terminal
additions became telescoped backward into shortened earlier stages. Thus the series of
ontogenetic stages was considered to he a condensation of successive ancestral adult
stages. Ernst Haeckel codified this theory as the Biogenetic Law. He recognized that
developmental features could he modified out of sequence, hut accommodated these
exceptions as merely inconvenient falsifications of the true history.

In a fascinating digression Gould reviews the profound effect that Haeckel's theory
had on such diverse areas as criminal anthropology, racism, theories of child development,
and Freudian psychoanalysis.

Early in this century the recapitulation theory- was gradually discarded not through
disproof (it accommodated all exceptions) hut because embryologists turned from a
search for ancestors to a search for the causal mechanisms of development, for which
the theory was irrelevant. Mendelian genetics showed that the determinants of heredity
could act not just terminally hut at all stages of development, and that genes often act
hy controlling the lates of processes. Recapitulation was abandoned as a universal prin-
ciple in favor of a general model in which it was only one of several modes of evolutionary
change resulting from shifts in the timing of developmental events.

152

THE WILSON BULLETIN • VoL 90, No. 1, March 1978

In Embryos and Ancestors (originally published as Embryology and Evolution) Gavin
de Beer described 8 categories of Heterochrony, or evolutionary changes in the timing of
developmental events. Gould argues that 4 of these involve the introduction of new
characters rather than changes in timing, but that 4 others are true examples of
heterochrony. Keorganizing de Beer’s ideas he presents a new theory of heterochrony
distinguishing between developmental processes and their morphological results. The
])rocesses are (1) Acceleration: the development of a feature is shifted to an earlier
stage in ontogeny, and (2) Retardation: a displacement to a later stage compared to
ancestral ontogenies. There are also 2 possible results. (^1) Recapitulation is the
repetition of ancestral adult stages in the descendent’s embryonic or juvenile phase.
It may arise either by the acceleration of a feature, or by prolongation of the life
history. (2) Paedomorphosis is the retention of ancestral juvenile characteristics by
later developmental stages of descendents. There are 2 distinct types of paedomorphosis.
Progenesis results from the precocious sexual maturation of a form with otherwise
juvenile morphology (often including small size). Neoteny results from the retardation
of somatic development, though body size may equal or exceed that of ancestral forms.
Examples of progenesis are drawn from various invertebrates; neotenic vertebrates in-
clude some urodeles, humans, and ratite birds. Gould, like many previous authors, re-
defines various terms. This source of possible confusion is clarified by an excellent
glossary in which many terms are not only defined, but changes in their meaning are
traced.

A unique aspect of Gould’s theory is the attempt to tie reproductive strategies to
environmental conditions. He suggests that paedogenesis allows adaptation to different
sets of environmental constraints on population growth. Progenesis is an adaptive re-
sponse to r-selective regimes, and neoteny to K-selection. Under conditions where
population size is not density restricted, r-selection predominates. Where populations
lie well below the carrying capacity of the environment, selection will favor strategies
that maximize growth, at least until the carrying capacity is reached. Such populations
tend to fluctuate widely in response to frequently unpredictable variations in the en-
vironment. Progenesis is effective here because early maturation and rapid develop-
ment favor rapid population increase. K-selection operates in more stable environments
to maintain relatively constant populations at or near the carrving capacity of the en-
vironment. K-selected organisms typically produce few young, but these have a relatively
hieh probability of survival to reproductive age. Populations are density limited, and
there is strong selection pressure for individual success. The argument is developed
by Gould mainly in terms of human evolution. Humans are shown to be neotenic,
with attendant long gestation and dependency periods.

The main problem here is that the validity of Gould’s thesis ultimately depends on
the validity of the existing theorv’ of r and K selection. Though wddely used as a
pedagogic model, the theory embodies many difficulties and may be too simplistic to
have the generality that Gould requires. The question was recently reviewed by Stephen
C. Stearns (Ann. Rev. Ecol. Syst. 8:145-71. 1977) wdio concluded that we still lack a
general and reliable theor>' of life historv' evolution.

Finally. Gould makes brief mention of recent biochemical studies showing that humans
and chimpanzees are nearly identical in structural genes, and that their profound
('rganismal differences must be due mainly to changes in genetic regulatorv' mechanisms.
He concludes that these proposed changes are heterochronic in nature.

Although this book contains few specifically ornithological references, its provocative
ideas should be of interest to all workers concerned with the evolution and ecological
relationships of birds. — Robekt J. Raikow.

March 1978 • ORNITIIOLOGICAL LITERATURE

153

Optical Signals: Animal Communication and Light. By Jack P. Hailman. Indiana
University Press, Bloomington, Ind., 1977 : 362 pp. $15.00. — A theoretical analysis of
visual communication in animals, including both markings and movements. Although
there are many examples dealing with birds, the emphasis is on the nature of the process
of communication rather than on any particular group of animals. — R.J.R.

North American Bird Songs: A World of Music. By Poul Bondesen. Scandinavian
Science Press Ltd., 2930 Klampenhorg, Denmark, 1977: 254 pp., over 225 sound spectro-
grams. $15.50. — This hook has some outstanding merits, some features difficult to eval-
uate, and some minor defects. The defects are to he expected, considering the complex-
ity and difficulty of the task the author set himself, which was to make a detailed study
of one large geographical segment of the world of bird music — a segment distant from his
own country — and to employ a language other than his own in expressing his results.
The courage and ability shown in this undertaking justify a certain patience on the part
of the reader. This is the more true as the task is unprecedented. A. A. Saunders (the
incomparable) did something like it for the Eastern U.S., hut before the days of sound
spectrograms. A. V. Arlton, in his Songs and Other Sounds of Birds (lithographed only),
put the vocalizations of 170 species into musical notation with some verbal description
and analysis, but most ornithologists are unable to profit from musical notations for songs.

The 290 species dealt with in this book are a large part of the passerine birds in the
continental U.S. and Canada, but do not include any nearctic species found exclusively
South of the Mexiean border. The spectrograms are each given more page space than
those in Birds of North America by Chandler Robbins et ah, hut they are mostly made
with a wide-hand machine and hence, while giving admirable temporal resolution, are
often less helpful as to frequency than the tiny ones in the other hook. One feels this
especially with clear-toned species, e.g., chickadees, Bachman’s Sparrow, or Hermit
Thrush. I prefer the narrow-band filter, but the ideal (apart from expense) would he
to he use both for each species.

The introductory essays on analyzing, describing, and reproducing songs and on song
behavior are helpful. There is a list of gramophone records, an index (of birds only),
a good glossary or “vocabulary,” a short but well chosen bibliography, and 439 items of
literature cited. In addition to his own careful analyses of songs, Bondesen quotes freely
from other authors. This helps to make the book a useful reference work. The verbal
descriptions and spectrograms admirably supplement one another.

Certain unidiomatic phrases and eccentric word usages may give readers trouble. Here
are perhaps the most serious: staccato fp. 233) for “short or long figures having an
almost eonstant frequency” (italics added); monotone (pp. 17, 42), for monotonous
repetition of “figures, motifs, or phrases” (which may internally have wide frequency
contrasts); phrase ( pp. 40, 70, 232) for what some writers call a song; continuous ( p.
41) for a sound seijuence “not consistently divided into phrases” (rather than for the
absence of substantial pauses between successive utterances). The above 4 words are
crucial for the author’s purposes, but only staccato and phrase are defined in the vocabu-
lary. I emphatically agree with the hint that “exclusion,” or the omission, now and
then, of a part cf a bird’s song-pattern or phrase contributes to variety, the achievement
of contrast.

Features not easily evaluated are: the classification or “key” for songs — rather dif-

ferent from the one Saunders provided — which Bondesen offers as an aid to identifica-
tion; also the arrangement of species in an order determined by the key rather than in
a standard taxonomic sequence. Thus Bachman’s Sparrow and the Blue-winged Warbler

151.

THE WILSON BULLETIN • Vol. 90, No. 1, March 1978

appear on the same page. The intricate and ingenious key is not easy to grasp, and
oj)inions as to its usefulness may differ. Singers are divided into 3 basic groups: I, II,
and III. or the Starling, Warhler, and Thrush groups. Some species have songs falling
into more than one group. More than two-thirds are put into Group II, which might also
he termed the Warhler-Finch group, since besides parulids, it includes most of our
fringillids. Group III includes turdids, mimids, vireos, and a few fringillids. The groups
are divided and subdivided by various criteria.

Starling-type songs (30 spp.) are largely formless, lacking in “architectonic structure.”
Both Warbler and Tbrush songs are more or less structured and well-patterned. Warblers
and many finches sing their phrases with monotonous repetition (“bound composition”)
and (at least in Group IIBl with long pauses, compared to the Thrushlike singers, with
their nonmonctonous or “free” composition and short pauses or none. These facts, as the
author seems to hint (p. 17), conform at least partially to my thesis that birds tend to
avoid repetition unrelieved by pauses.

Bondesen refers to Dobson and Lemon (Nature, 257:26-28, 1975) who critically discuss
this “antimonotony” rule. The discussion helps to clarify the issue. Biological or be-
havioral generalizations usually require qualifications; some of these are (though in-
adequately) stated in my book. Born to Sing, which the author lists. What I meant hy
the “monotony threshold” was not a positive correlation of overall versatility ( size of
repertoire) with continuity of singing (high ratio of song or phrase lengths to lengths
of pauses), but rather a negative correlation between singing repetitively, or without
“immediate variety” (changes of tune not after minutes but after at most a few seconds)
and the absolute length of pauses between phrases. Monotony, for fast-living singing
animals, is avoided either by changing the music or by pauses long enough for memory
to fade. With this understanding both Bondesen ’s book and the Dobson-Lemon tables
support this rule, which I still claim as a discovery.

Even without considering the distinction between immediate and eventual variety, the
tables yield the following. The 39 species are considered in order of increasing repertoire-
sizes. Dividing them into 3 groups, ##1-13, 14-26, 27-39, and considering the pause
lengths in each group we have:

Spp. with very short pauses (less than 1.5

Group I
secs. ) 0

Group II
1

Group III
4

S])p. with short pauses (less than 3 secs.)

Spp. with long pauses (more than 6 secs.)

Average pause length:

6.39

4.58

3.70

Average repertoire size:

26+

This exhibits the “negative correlation between interval length and repertoire size” that
Dobson and Lemon concede. It supports my view that songbirds tend to escape monotony
either hy musical variety or by substantial (for a bird’s brain) lapses from singing. Tak-
ing the question of immediate versus eventual variety into account would, I predict,
further support this conclusion, though adecjuate data are lacking. Thus the Song Spar-
row and (,’ardinal, with long pauses, could be put into Group I as, for the monotony
(piestion, limited in variety. The Song Sparrow, though with many songs, sings quite
rej)ctitively in the short run. Anotlier (jualification : counting distinguishable patterns on
a spectrogram is not an accurate measure of variety. Tbere is also the depth of the con-
trasts. The Rose-brt‘asted Grosl)eak is listed as pausing for 10 seconds and yet highly
versatile, but to me the contrasts seem slight.

March 1978 • ORNITHOLOGICAL LITERATURE

155

Dohson and Lemon take too lightly the crucial evidence that individuals of some
species ( e.g.. the Eastern Peewee) sing at times with immediate variety and short pauses;
at other times without immediate variety hut with long pauses; and in no third way.
Three groups of Nightingale \^’rens ( Microcerculus marginatus ) of tropical America oc-
cur: in one every successive sound is on almost the same pitch and pauses are several
seconds; in another every sound is on a different pitch, with very short pauses; and in
a third the pitch changes about every other sound, with pauses of intermediate length.
The monctony threshold has not been disproved, whatever qualifications may be needed.
All three authors have done me a good turn by applying such careful consideration to
the topic.

To write in detail about hundreds of birds without mistakes is scarcely possible. Lark
Sparrows do not “always,” perhaps not even usually, sing in flight, and even the .Skylark
may sing from a perch. Borror is miscited < p- ”0* as writing that a Carolina Wren has a
singing-rate of 4-24 motifs per minute. \^'hat he correctly wrote was not motifs but songs.
Bondesen’s alteration underestimates by 3 times the amount of singing per minute. How-
ever, such mistakes seem relatively few.

My conclusion is that students of bird song need this l)ook. ^ e can be grateful that
the author dared to undertake such a difficult task. It is the outstanding, up to date re-
gional study of bird song. Although slightly too big for some jacket pockets, it is rea-
sonably convenient to carry in the field as an aid to identification. — Charles Hartsiiorne.

John Gould Bird Print Reproductions. By Gordon C. .Sauer. Privately printed;
order from Richland Enterprises, P.O. Box 7062, Kansas City, MO., 64113; 76 pp., paper
cover. S4.50. — This is not a collection of Gould’s prints. Instead, its purpose “is to
assist others in correctly identifying the origin of a particular Gould bird print. This
information should be especially valuable for interior decorators, print dealers, and the
many others who appreciate the beautiful colored prints of Gould’s birds.” — R.J.R.

H. Hudson, A Bibliography. By John R. Payne. Archon Books, The .‘^hoe String
Press, Inc., Hamden CT. 1977:248 pp., 1 phcto of U . H. Hudson. .S17.50.

The Biological and Taxonomic Status of the Mexican Duck. By John P. Hubbard.
Bulletin no. 16, New Mexico Dept, of Game and Eisb. 1977:56 pp. No price given. —
Based on an analysis of plumage and other characteristics, Hubliard concludes that the
Mexican Duck is a subspecies idiazi) of the Mallard ^ Anas platyrhynchos) and not a
separate species. The distribution and biological characteristics of the Mexican Duck
are discussed, and recommendations are made concerning future management policies. —
R.J.R.

156

THE WILSON lU'LLETIN • Vol. 90, No. 1, March 1978

A Classification of the Tykant Flycatciieks (Tyhannidae) . By Melvin A. Traylor,
jr.. Bull. Mus. Comp. Zool., Harvard Univ., Vol. 148, No. 4: 129-184, 1977. No price
given. — This i)aper is an explanation of the new classification of the Tyrannidae that
Traylor has prepared for volume 8 of Peters’ Check-list of Birds of the W orld. It is
based on examination and measurements of study skins, as well as on distribution,
behavior, and other information from the literature. The most important references used
are Peter Ames’ study of the syrinx (Peabody Mus. Nat. Hist. Bull. 37, 1971) and
Stuart Waiters’ unpublished thesis on the cranial osteology of the Tyannoidea (Louisiana
State University, 1965). The last classification of the whole family was published by
Hellmayr in 1927. The family includes all of Hellmayr’s genera plus 5 genera formerly
classified in the Cotingidae, and Corythopis from the former family Conopophagidae.
The 7 subfamilies of Hellmayr have been reduced to 3, the Elaeniinae, Fluvicolinae, and
Tyranninae. The genera in each subfamily are listed in order from generalized to
specialized types. Since this is often vague and arbitrary Traylor has simply retained
Hellmayr’s sequence (in reverse) unless there are compelling reasons for change.
Thirty-six of Hellmayr’s genera are synonymized, one is resurrected, and a new genus
Zimmerius is described for 5 species formerly in Tyranniscus. The classifications of
previous authors are contrasted, and the reasoning behind the new classification is given
in lengthy detail.

An attempt is made to analyze the phylogeny of the Tyrannidae by cladistic analysis
hut is only marginally successful. Most characters are so variable and subject to such
frequent convergence that their primitive and derived states could not be established.
Only Warter’s data on the skull are so analyzed. The methods used to determine
morphocline polarities are described only vaguely. They appear to be based on the
idea that character states widely distributed among more than one subfamily are

primitive, and on an intuitive assumption that primitive flycatchers were small arboreal
forms, nearest the present Elaeniinae. The phylogenetic diagram (Fig. 7, p. 173) has
a low degree of resolution. Four separate lineages arise from a hypothetical common
ancestor, one to the Becards and Tityras, and the others to the 3 subfamilies. The
lineages to the Tyrannidae and Fluvicolinae each branch into two groupings. This
low resolution clearly demonstrates the need for additional studies with other kinds of
data, a point that Traylor emphasizes several times. Nevertheless, there is an encourag-
ing degree of correspondence between Ames’, Warter’s, and Traylor’s groupings. This
indicates that the arrangements, as far as they go, are soundly based.

A new classification of the Tyrannidae has been sorely needed because of the

abundance of new information that has accumulated in the half centun- following
Hellmayr’s work. Traylor’s thorough study fills this need admirably to the degree that

present knowledge permits, and will serve as a solid basis for the new studies that we

may hope it will stimulate. — Robert J. Raikow.

ORNITHOLOGICAL NEWS

Aaron M. Bagg Student Membership Awards

Student Membership Awards in the Wilson Ornithological Society have been made
available through funds generously donated in memory of the late Aaron M. Bagg, for-
mer president of the Society. The Student Membership Committee has designated the
award recipients for 1978 as follows: Gary R. Alten, California State Polytechnic Uni-
versity; Darrel C. Boone, University of Maryland; Michael C. Delesantro, New Mexico
State University; Claire L. Filemyr, Virginia Commonwealth University; Wayne Hoff-
man, University of South Florida; Anthony H. James, San Francisco State College;
Samuel F. Jojola, New Mexico State University; Thomas R. Kemp, University of
Toledo; Sandra J. Korowotny, Texas A&M University; Marc D. Longwood, California
State University, Sacramento; Selby R. Mohr, California State University; Erica Nol,
University of Guelph; Christopher M. Rogers, University of Wisconsin, Milwaukee;
Kathryn J. Schneider, Princeton University; Theodore R. Simons, University of Wash-
ington; Thomas W. Smith, Jackson, Kentucky; Shirley J. Thompson, University of To-
ronto; Melinda J. Welton, University of Connecticut.- — James R. Karr, Chairman, Student
Membership Committee.

1978 Annual Meeting

The 59th annual meeting of The Wdlson Ornithologieal Society will he held at Jackson’s
Mill, West Virginia, on 4-7 May 1978. The meeting will be hosted by the Brooks Bird
Club, the Department of Wildlife Biology of West Virginia University, and West Vir-
ginia University.

A special feature of the meeting will he a symposium titled, “Resource Use Strategies
in Birds,” to he held on the afternoon of Friday, May 5th. The symposium is organized
by Dr. Elliot J. Trainer. The chairman of the Local Committee is Dr. Robert Whitmore,
Division of Forestr\-, West Virginia University, Morgantown, WV 26506.

Notice of a Possible Numerical Coding System for All Bird Species

Increased use of computers to store and process data about birds has precipitated a
number of problems. One such problem is that of identifying the species (or higher
taxonomic unit) under consideration. Many local ornithological organizations have
solved this problem by identifying each species in their area with a unique code
number ( AOU number, for example). The proliferation of local systems could he
avoided by development of an internationally recognized coding system for all bird
species. The advantages of a standard system include international compatibility of
records and facilitation of exchange of data and literature among countries.

Any new system must he taxonomically based, flexible enough to accommodate new
species and taxonomic revisions, and he expandable for those interested in suhspecific
classifieation. For purposes of discussion, the following system is proposed. A world-
wide numerical system will he based on Morony, Bock, and Farrand (Reference List
of the Birds of the World, 1975, AMNH) with six digit numbers identifying each species.

Thus, each genus has numbers reserved in advance for up to 99 species and searches
of data can he made rapidly and efficiently by computer for any taxonomic level. In-
dividuals wanting suhspecific identifications can simply add one or two digits to the
6-digit base.

157

The Wilson Bulletin

Editor* JtROME A. Jackson

Department of Zoology
P.O. Drawer Z
Mississippi State University
Mississippi State, MS 39762

Editorial Assistants Bette J. Schardien Patricia Ramey

C. Dwight Cooley Martha Hays

Renne R. Lohoefener

Review Editor Robert Raikow Color Plate Editor William A. Lunk

Department of Life Sciences 865 North Wagner Road

University of Pittsburgh Ann Arbor, MI 48103

Pittsburgh, PA 15213

Suggestions to Authors

See Wilson Bulletin, 87:144, 1975 for more detailed “Suggestions to Authors.”
Manuscripts intended for publication in The Wilson Bulletin should be submitted in dupb-
cate, neatly typewritten, double-spaced, with at least 3 cm margins, and on one side only
of good quality white paper. Do not submit xerographic copies that are made on slick,
heavy' paper. Tables should be typed on separate sheets, and should be narrow and deep
rather than wide and shallow. Follow the AOU Check-list (Fifth Edition, 1957) and
the 32nd Supplement (Auk, 90:411-419, 1973), insofar as scientific names of U.S.
and Canadian birds are concerned. Summaries of major papers should be brief but
quotable. Where fewer than 5 papers are cited, the citations may be included in the text.
All citations in “General Notes” should be included in the text. Follow carefully the style
used in this issue in listing the literature cited; otherwise, follow the “CBE Style Manual”
(1972, AIBS). Photographs for illustrations should have good contrast and be on gloss
paper. Submit prints unmounted and attach to each a brief but adequate legend. Do not
write heavily on the backs of photographs. Diagrams and line drawings should be in black
ink and their lettering large enough to permit reduction. Original figures or photographs
submitted must be smaller than 22 X 28 cm. Alterations in copy after the type has been
set must be charged to the author.

Notice of Change of Address

If your address changes, notify the Society immediately. Send your complete new
address to the Treasurer, Ernest E. Hoover, 1044 Webster St., N.W., Grand Rapids,
Michigan 49504. He will notify the printer.

The permanent mailing address of the Wilson Ornithological Society is: c/o The

Museum of Zoology, The University of Michigan, Ann Arbor, Michigan 48104. Persons
having business with any of the officers may address them at their various addresses
given on the back of the front cover, and all matters pertaining to the Bulletin should be
sent directly to the Editor.

* See Ornithological News, p. 158, for address for ms submission.

CONTENTS

GROWTH AND SURVIVAL OF YOUNG FLORIDA SCRUB JAYS Glen E. Woolfeuden

VERTICAL DISTRIBUTION OF BIRDS IN A LOUISIANA BOTTOMLAND HARDWOOD FOREST

James G. Dickson and Robert E. Noble

AGRICULTURAL IMPACT OF A WINTER POPULATION OF BLACKBIRDS AND STARLINGS

Richard A. Dolbeer, Paul P. W' oronecki, Allen R. Stickley, Jr., and Stephen B. W hite

BREEDING BEHAVIOR OF THE LOUISIANA HERON James A. Rodgers, Jr.

STATUS AND NUMERICAL FLUCTUATIONS OF SOME NORTH AMERICAN WADERS ALONG THE

SURINAM COAST Arie L. Spaans

FEEDING OF NESTLING AND FLEDGLING EASTERN BLUEBIRDS Benedict C. Pinkowski

DIFFERENTIAL USE OF FRESH WATER ENVIRONMENTS BY WINTERING WATERFOWL OF COASTAL
TEXAS Donald H. W'hite and Douglas James

REPRODUCTIVE SUCCESS AND FORAGING BEHAVIOR OF THE OSPREY AT SEAHORSE KEY, FLORIDA

Robert C. Szaro

GENERAL NOTES

CHANGING AVIAN COMMUNITY STRUCTURE DURING EARLY POST-FIRE SUCCESSION IN THE

SIERRA NEVADA Carl E. Bock, Martin Raphael, and Jane H. Bock

NOTES ON THE DISTRIBUTION OF BIRDS IN SONORA, MEXICO

Stephen M. Russell and Donald W . Lamm

EGG CARRYING BY WOOD DUCK

Robert W. Strader, Richard Di Giulio, and Robert B. Hamilton

EVIDENCE OF BROOD ADOPTION BY RUFFED GROUSE Stephen J. MaXSOn

MARSH HAWKS FOLLOW HUNTING RED FOX LeRoy W. Bandy and Barbara Bandy

PREDATION ECOLOGY OF COEXISTING GREAT HORNED AND BARN OWLS Seri G. Rudolph

HOST RECORDS FOR THE STRIPED CUCKOO FROM COSTA RICA

Lloyd F. Kiff and Andrew Williams

ANT-FOLLOWING BIRDS IN SOUTH AMERICAN SUBTROPICAL FORESTS

Michael Gochfeld and Guy Tudor

FISHING BEHAVIOR OF BLACK AND TURKEY VULTURES

Jerome A. Jackson, Irvine D. Prather, Richard N. Conner, and Sheila Parness Gaby
A NEW HYBRID WARBLER COMBINATION Richard C. Banks and James Baird

ORNITHOLOGICAL LITERATURE

ORNITHOLOGICAL NEWS

18, 30,

112

119
123 ^

131

132

133

134

138

139 /

141 1<
143 \\

145

157 I

REQUESTS

FOR

ASSISTANCE

TfieWlsonBulleftn

PUBLISHED BY THE WILSON ORNITHOLOGICAL SOCIETY

VOL. 90, NO. 2 JUNE 1978 ^US.

library

SEP 1 1978

HARVARD

Ur^iVBERaiTY

The Wilson Ornithological Society
Founded December 3, 1888

Named after ALEXANDER WILSON, the first American Ornithologist.

President — Douglas A. James, Department of Zodlogy, University of Arkansas, Fayetteville,
Arkansas 72703.

First Vice-President — George A. Hall, Department of Chemistry, West Virginia Univer-
sity, Morgantown, W. Va. 26506.

Second Vice-President — Abbot S. Gaunt, Department of Zoology, Ohio State University,
Columbus, Ohio 43210.

Editor — Jerome A. Jackson, Department of Biological Sciences, P.O. Drawer Z, Missis-
sippi State University, Mississippi State, Mississippi 39762. (See Ornithological
News, p. 308) .

Secretary — Curtis S. Adkisson, Department of Biology, Virginia Polytechnic Institute
and State University, Blacksburg, Virginia 24061.

Treasurer — Ernest E. Hoover, 1044 Webster St., N.W., Grand Rapids, Michigan 49504.

Elected Council Members — James R. Karr (term expires 1979) ; Clait E. Braun (term
expires 1980); Sidney A. Gauthreaux, Jr. (term expires 1981).

Membership dues per calendar year are: Active, $10.00; Sustaining, $15.00;

Life memberships, $200 (payable in four installments).

The Wilson Bulletin is sent to all members not in arrears for dues.

The Wilson Bulletin

The official organ of the Wilson Ornithological Society, published quarterly, in March, June, September,
and December. The subscription price, both in the United States and elsewhere, is $15.00 per year. Single
copies, $4.00. Subscriptions, changes of address and claims for undelivered copies should be sent to the
Treasurer. Most back issues of the Bulletin are available and may be ordered from the Treasurer. Special
prices will be quoted for quantity orders.

All articles and communications for publications, books and publications for reviews should be addressed to
the Editor. Exchanges should be addressed to The Josselyn Van Tyne Memorial Library, Museum of Zoology,
Ann Arbor, Michigan. Known office of publication : Department of Zoology, Mississippi State University."
Mississippi State, Mississippi 37962.

Second class postage paid at Mississippi State, Mississippi and at additional mailing office.

Allen Press, Inc., Lawrence, Kansas 66044

1.,

Head patterns of Mexican Piculus. From top to bottom: 1 . “Typical” auricularis. 2. Variant auricularis
with red flecks surrounding pileum. 3. Rubiginosus (either yucatanensis or maximus), with conspicuous s |
red band surrounding pileum. 4. Variant aeruginosus with thin red line from lore to above the eye. I
5. "Typical” aeruginosus with red restricted to the nape. Watercolor by George M. Sutton. !1

THE WILSON BULLETIN

A QUARTERLY MAGAZINE OE ORNITHOLOGY
Published by the Wilson Ornithological Society

VoL. 90, No. 2 June 1978 Pages 159-334

Wilson Bull., 90(2), 1978, pp. 159-181

A REVISION OF THE MEXICAN PICULUS
(PICIDAEj COMPLEX

Luis E. Baptista

The neotropical woodpeckers of the genus Piculus are closely related
to the flickers iColaptes) (Short 1972 j. Piculus species range from Mexico
to southern Brazil, Paraguay, Peru ( Ridgway 1914 j and Argentina ( Salvin
and Godman 1892 ) . Peters ( 1948 j lists 46 taxa ( 9 species and their sub-
species ) of which 20 are subspecies of Piculus rubiginosus, the most widely
distributed species. The latter ranges from southern Veracruz to the north-
western provinces of Jujuy, Salta, and Tucuman in Argentina (Peters
1948).

Compared with other picids, this genus is generally poorly represented in
museum collections. It is possible that they are not as rare as they seem,
hut being rather silent and secretive birds and difficult to distinguish from the
associated vegetation due to their cryptic green coloration, are easily passed
unnoticed by collectors in the field.

A difference of opinion exists among taxonomists regarding the status
of several of the Mexican forms. Two species complexes are recognized in
the Mexican check-list (Miller et al. 1957). The Piculus auricularis complex
is reported by these authors as consisting of 2 subspecies: sonoriensis

known only from the type series of 3 birds taken at Rancho Santa Barbara,
Sonora, and the nominate race auricularis recorded as ranging from Sinaloa
south to Guerrero. They point out the uncertain status of the form sonorien-
sis, stating that additional material is needed to substantiate it. In their
treatment, the Piculus rubiginosus complex in Mexico is subdivided into 3
subspecies: aeruginosus, maximus, and yucatanensis. Other authors have
treated aeruginosus as a full species ( Ridgway 1914, Peters 1948, Sutton
1951, Blake 1953, Peterson and Chalif 1973, Gehlbach et al. 1976). Wetmore
(1941) questioned the status of maximus stating that “the systematic under-
standing of the species rubiginosus is at present unsatisfactory. In Mexico
and Central America these birds seem subject to much individual variation,
and I am inclined to believe that too many races have been proposed.” In
discussing the status of yucatanensis. Miller et al. (1957) point out that “all

159

ICO

THE W ILSON lU LLETLN • VoL 90, \o. 2, June 1978

specimens from high altitudes in the interior of south central Mexico re-
(juire critical re-examination.”

This study treats morphological variation in the Mexican Piculus complex
and evaluates the status of the described forms. Piculus auricularis so far
has proved to he allopatric in its geographical distribution with the rubigi-
nosus complex. Several authors (Van Rossem and Hachisuka 1937, Sutton
1951, 1953) have pointed out the similarities existing between these species
suggesting that they might actually be conspecific. Is the genus Piculus,
therefore, represented by 1, 2, or 3 species in Mexico?

MATERIALS AND METHODS

Museum samples of Piculus ^ rubiginosus \ aeruginosus and Piculus rubiginosus
yucatanensis allopatric throughout most of their ranges were examined critically.
Specimen material from central Veracruz was scrutinized for evidence of free inter-
breeding and intergradation which if present would justify the treatment of aeruginosus
as a race of the rubiginosus complex, and if absent would warrant recognizing them as
2 distinct species. Material representing populations of Piculus rubiginosus in southern
Mexico, namely in Oaxaca, Chiapas, and southern Veracruz, was analyzed, and the
validity of the races yucatanensis and maximus evaluated.

For simplicity's sake the races of Piculus rubiginosus are referred to throughout this
paper by their subspecific names aeruginosus, yucatanensis, and maximus. Similarly
the races of Piculus auricularis are referred to as auricularis and sonoriensis. Some 238
skins from Mexico were considered in this study: 68 of Piculus auricularis, 81 of

aeruginosus, 33 of yucatanensis. and 56 of maximus. In addition, 13 skins of yucatanensis
from Honduras were examined. Localities from which samples were examined are
illustrated in Fig. 1.

Measurements were taken of bill length (from nostril), tarsus, and chord of wing
as described by Baldwin et al. (1931). Except when otherwise mentioned, one-tailed
t-tests were conducted to test for differences between adjacent samples.

ECOLOGY AND LIFE HISTORY NOTES

Short ( 1973 I has called attention to the paucity of information regarding
the biology of Piculus species. I have, therefore, brought together the scanty
literature on the natural history of the Mexican forms, which, hopefully,
may prove useful to investigators intending to carry on field work on this
group.

In the northernmost part of its range in Sonora, Piculus auricularis so-
noriensis is known only from its type locality in the Upper Sonoran Zone
at 1500 m elevation (Van Rossem 1945). Vegetation found in this life zone
is discussed in Orr (1966:274). Elsewhere in its range it is a bird of pine-
oak, pure oak, or oak-tropical deciduous forest. Specimens I examined
were taken in pine-oak as high as 1970 m in Babizos, Sinaloa, to as low as
900 m in oak woodland near lepic. Xayarit. To my knowledge it has never

Baptista • REVISION OF MEXICAN FICULUS

161

Fig, 1. Map of Mexico showing localities from which samples of Mexican Ficulus
were examined.

been taken in the thorn scrub of coastal western Mexico. Much forest land
has been cleared for cultivation; if abandoned, fallow land is invariably
invaded by thorn scrub so that Ficulus habitat is fast decreasing. In the
southernmost part of its range in the Sierra de Miahuatlan, Oaxaca, the
nominate auricularis has been taken in a greater variety of habitats such as
boreal forest at 2650 m, cloud forest at 1500-2100 m, oak-tropical deciduous
forest at 770 m, and humid tropical evergreen forest at 740 m. Habitats in
these collecting localities are described in detail by Rowley ( 1966 ) . The
species probably occupies similar habitat in Guerrero (see vegetation map in
Leopold 1959 ) .

The Mexican Ficulus^ like others of this genus, in general are rather silent
and secretive birds. The chief distinguishing feature between Colaptes and
Ficulus is the latter’s silent nature (Short pers. comm.). There seems to be
nothing in the literature concerning the vocalizations of Ficulus auricularis
or its nesting habits. Scbaldach ( 1965 ) comments on the “apparent rarity”

162

THE WILSON BULLETIN • Vol. 90, No. 2, June 1978

of this species, suggesting that this may be due to its being very widely dis-
tributed, each individual covering a larger territory than those held by other
woodpecker species.

Piculus rubiginosus aeruginosus has been taken in pine-oak 1500-2100 m
in Nuevo Leon and in tropical forest in Gomes Farias, Tamaulipas (Sutton
and Pettingill 1942). C. C. Lamb (unpubl. field notes, Moore Laboratory
of Zoology) took a series of this form 24 km southwest of Linares, Nuevo
Leon, where the habitat consisted of white and live oaks with mesquite trees
forming an understory. This bird was also found to be fairly common in
the canyon bottoms of the Sierra de Tamaulipas below 600 m and occasionally
in pine-oak woods at 900 m (Martin et al. 1954) . At La Joya de Salas, Tamau-
lipas, it was mainly a bird of deciduous woodland (Robins and Heed 1946).
C. C. Lamb (unpubl. field notes) took this form on cottonwoods at the Rio
Corona, 25 km north of Ciudad Victoria, Tamaulipas. There the vegetation
consisted of sizeable trees intermixed with various cacti and mesquite trees.
In the southern part of its range at Huachinango, 360 m, near the Puebla-
Veracruz line. Lamb (field notes) collected this subspecies in “dense jungle.”

The subspecies P. r. yucatanensis is a bird of mesic conditions (Wetmore
1943:222), similar to aeruginosus in Puebla and Veracruz. In the rain-
forests of the Catemaco Basin of Veracruz, it was found to be mostly a forest
edge species (Edwards and Tashian 1959). At Cordoba, Veracruz, 270 m.
Lamb (field notes) took this subspecies in heavily wooded mountains, de-
scribed as a tangle of vines and bushes.

The subspecies P. r. maximus inhabits more open forest, such as the pine-
oak and riparian tropical hardwood of the Monseratte Plateau ( Edwards and
Lea 1955). At Finca Cacahuatl, Chiapas, 19 km east of Tapanatepec,
Oaxaca, Lamb (field notes) found this species in pine and oak forests at
808 m.

Dickey and Van Rossem (1938) found that P. r. yucatanensis fed exclu-
sively on insects. They described its foraging behavior as working slowly
up a tree and gently prying into crevices. Wetmore (1968) studied the spe-
cies in Panama where he found that the diet included fruit. He observed a
male eating a large blackberry. C. C. Lamb (field notes) observed a female P.
auricularis eating berries off a madrone tree at Babizos, Sinaloa.

Piculus rubiginosus is very flicker-like in many of its habits sucb as
in vocalizations (Dickey and Van Rossem 1938, Sutton 1951, 1953, Blake 1953,
J. S. Rowley pers. comm.), its manner of perching (Sutton 1951), and its
courtship behavior as observed by Sutton (1942, 1953). The last author re-
ported seeing 3 or 4 birds together with spread wings and tail, bobbing and
bowing to each other while calling excitedly. This activity was interrupted
with brief periods of statuesque motionlessness. Group displaying is also

Baptista • REVISION OF MEXICAN PICULUS

163

known in at least 2 species of flickers, the Andean Flicker (Colaptes rupicola)
and the Campo Flicker (C. campestris) (Short 1972).

Nest holes have been found from 3.6-9 m off the ground in dead and live
trees (Sutton 1953, Skutch 1969, Rowley pers. comm.j. Clutch size appears
to be 4 in aeruginosus (Robins and Heed 1951), maximus (Rowley pers.
comm.), uropygialis (Skutch 1956), and trinitatis (Belcher and Smooker
1936). Rowley took a set of 4 eggs of maximus at Cerro Raul, Oaxaca, 1300
m, in a nest in the cavity of a dead tree, 15 m above the ground by a creek
(HC 21387). Lloyd Kiff kindly provided egg measurements (mm) which
are as follows: 23.91 X 19.10, 24.39 X 19.20, 24.60 X 18.92, 23.43 X 18.29.

The only detailed observations on nesting behavior of P. rubiginosus
are those of Skutch (1943, 1948, 1956, 1969) which are here summarized.
Non-nesting individuals roosted solitarily in holes. A male was observed
joining a female in her roosting hole which was converted to a nest. Four
eggs were observed resting on clean chips. Both sexes alternated on the
eggs during the day, and the male incubated by night. Three hatched to-
gether, and 1 the next day. The young were pink-skinned and naked on
hatching. Pinfeathers were first observed at 8 days. The parents removed
waste matter from the nest only until the young were old enough to take
food from the nest entrance at approximately day 21. Both parents fed the
young. Only one of a brood of 4 observed survived. This was a female
nestling which was flushed from the nest at day 24 and flew off. An adult
male used the nest hole for roosting after the young had fledged. Young in
their first plumage already had markings of adults of their own sex.

THE STATUS OF Piculus auHcularis sonoriensis

The subspecies sonoriensis was described by Van Rossem and Hacbisuka
(1937) as grayer on pileum and back than the nominate, with “the upper
back between the nape and dorsum prominently barred with grayish white.”
The type locality given was Rancho Santa Barbara, 1500 m, 31.7 km north-
east of Guirocoba, situated at latitude 27° 16' and longitude 108° 35'
(Van Rossem 1945).

The type and a topotype examined and compared with material from other
parts of Mexico has led me to conclude that the gray coloration in the type
description is of an adventitious nature rather than of genetic origin. The
olive-green on the backs of Piculus is the result of the combined effect of 2
pigments distributed through 2 different components of the feathers. Be-
neath a dissecting microscope ( lOX ) black pigment may be seen in the bar-
bules, and yellow pigment in the rachis and barbs. Graying may be the result
of any of a variety of causes such as feather wear, fading of yellow pigment
due to exposure to light, to diet as suggested for Colaptes by Short (1965),

164

THE WILSON BULLETIN • Fo/. 90, No. 2, June 1978

or the leaching effect of tannin from the barks of trees on which they forage
as suggested for Picoides stricklandi by Davis 11965:573). Whatever the
cause, graying as described for sonoriensis was also found in samples of
auricularis from Nayarit and Oaxaca, as well as in several rubiginosus. The
latter appeared darker due to heavier deposition of melanin pigment.

The second character mentioned by Van Rossem and Hachisuka, the barred
upper back, was also found in series taken throughout the rest of the species’
range. Moreover, a specimen taken in nearby Mount Mohinora in Chihuahua
in October, in fresh fall plumage, was olive-green on the back as in auricularis
from Sinaloa and Nayarit. The specimens from Sonora are, therefore, not
subspecifically distinct at least from specimens taken in neighboring states
such as Chihuahua, Sinaloa, and Nayarit as stated by these authors.

Birds from the northern states (Sonora to Colima) are larger than those
from southern states (Guerrero and Oaxaca I (Figs. 2-4). Northern males
have longer wings and bills and northern females have longer wings than
do southern birds. Males of sonoriensis average 125 mm in wing length,
and males of auricularis average 119.5 mm (one-sided p < 0.00005). Mean
wing length of female sonoriensis was 122.2 mm and that for female auric-
ularis was 119.2 mm (one-sided p < 0.002). Bills in male sonoriensis
averaged 20.3 mm and those in male auricularis averaged 18.6 mm ( one-sided
p < 0.0005 ) .

The southern samples were also darker than the northern. This darkening
was not dramatic and taken alone does not constitute a good distinguishing
character. Hargitt ( 1890:183) gives measurements of the type specimen of
the nominate auricularis taken in Xautipa, Guerrero, as follows: culmen

23.5 mm, wing 117 mm, tail 72 mm, and tarsus 20.8 mm. These figures
suggest that Hargitt’s type may be placed with the southern samples. The
northern samples representing material from Sonora to Colima may be
recognized as a race distinct from the nominate being slightly lighter in
coloration, and larger in some morphometric characters discussed earlier.
These must be known as Piculus auricularis sonoriensis Van Rossem and
Hachisuka. The nominate race is, therefore, restricted to Guerrero and
Oaxaca as far south as Pochutla on the road to Puerto Angel from Oaxaca
City.

Schaldach (1963) recorded the first specimens of this species for the state
of Colima. The bird from Chihuahua reported herein ( H.C. $ 4728) is be-
lieved to be the first record for that state. The material from Pochutla,
Oaxaca ( DM $ 25046, DM 9 25045, DM 9 38824 1 extends the range of
the species from its former southern range in Guerrero (Miller et al. 1957).
To date no specimens have been reported for the state of Michoacan.

The pine-oak forests of Jalisco continue on into Michoacan. However, a

Baptista • REVISION OF MEXICAN PICULUS

165

belt of arid tropical scrub separates the Michoacan pine-oak forests from
those in Guerrero (see map in Leopold 1959:16). Piculus a. sonoriensis is
to be expected in Michoacan, but these populations are probably separated
from those in the Sierra Madre del Sur (auricularis) by the xeric belt. It is
conceivable, however, that if specimens from Michoacan are obtained, these
may prove to he intermediate in size between the forms, in which case
sonoriensis should be merged in auricularis.

Piculus auricularis as a subspecies of P. rubiginosus?

I found no evidence of interbreeding between P. auricularis and P. r.
yucatanensis from which it seems to he separated by the Isthmus of
Tehuantepec. Although previous authors have described auricularis as be-
ing entirely devoid of red on the crown, there were vestiges of red pigment
on the tips of the crown feathers bordering the pileum in varying amounts on
some specimens that I examined from throughout the species’ range. Red
spotting on the crown was described in a juvenile by Ridgway (1914) hut
is not necessarily limited to that age class. I interpret this as a recapitula-
tion of an ancestral character indicating that auricularis is a derivative of
the rubiginosus group to which it is similar in many other respects. An
analogous situation may he found in the conure Aratinga astec which shows
a tuft of orange feathering above the cere (Hardy 1966:66) suggesting a
common ancestor with the orange-fronted Aratinga canicularis. I have never
found a female of Piculus auricularis, however, showing any trace of red on
the crown in contrast to females of the rubiginosus group which always
have conspicuously red napes. The rubiginosus forms also have notably
darker crowns than do auricularis. The pileum is slate-gray in the former
and light-gray in the latter, although as a result of the color dine auricularis
from the Sierra de Miahuatlan, Oaxaca, approach yucatanensis with regard
to this character.

In some groups of birds, notably the parrots ( Psittaciformes) , small dif-
ferences in color or color patterns are important in social recognition and
may serve as effective isolating mechanisms between species I Hardy 1966,
1967). Experiments by Noble (1936) have demonstrated the importance
of the malar stripe in sexual recognition of the common flicker. Jerome
Jackson ( pers. comm.) blackened the red nape patch on a male Downy
Woodpecker ( Picoidcs pubescens ) . Its mate treated the disguised male as
another female and attacked it. Thus the presence ( as in Piculus rubiginosus )
or absence ( P. auricularis ) of a red nape patch may he an effective ethological
isolating mechanism between the two forms should they ever prove to
breed sympatrically.

Figs. 2^ reveal a decreasing size dine from sonoriensis to auricularis’.

166

THE WILS()x\ Bl LLETIN • To/. 90, Ao. 2, June 1978

BILL

I 1 1 1 1 1 1 1 1 1 1

15 17 19 21 23 25

m m

Fig. 2. Variation in bill length in Mexican Picul us. Horizontal lines denote ranges,
vertical lines means, with rectangles as 95% confidence intervals on each side of the
mean. Black rectangles denote males, clear rectangles signify females. Numbers in-
dicate sample sizes. Aeruginosus-n = northern aeruginosus from Nuevo Leon, San
Luis Potosi, and Tamaulipas, and aeruginosus-^ — birds from Puebla and Veracruz.

however, with regard to measurements of bill and tarsus, this does not con-
tinue into yucatanensis which shifts to the right, i.e. toward larger values.
The 2 complexes probably formed a continuous population down the west
coast of Mexico at one time and are now separated by a belt of tropical
deciduous forest (see map in Leopold 1959:16). Isolated from populations
south of the Isthmus, and thus without the genetic load of eastern Mexican
and Central American moister habitat populations, the accumulation of micro-
mutations has resulted in the evolution of the northern population into its
present form with overall lighter coloration (probably an adaptation to more
xeric conditions ) and in which selection against red on the crowns of females
has been complete and is almost so in males. The available, indirect evidence,
therefore, indicates that Piculus auricularis should be treated as a full species
distinct from ruhiginosus. This is in agreement with the Mexican check-list
(Miller et al. 1957).

THE TAXONOMIC STATUS OF PiculuS OeruginoSUS

Authors who have treated the form aeruginosus as a distinct species dis-
tinguished it from the ruhiginosus complex, and in particular from the sub-
species yucatanensis which replaces it geographically in central Veracruz, on
the basis of the following characters:

Baptista • REVISION OF MEXICAN PICULUS

167

WING

I I I I I I I

110 120 130 140

m m .

Fig. 3. Variation in wing length in Mexican Piculus. See Fig. 2 for explanation of
figure.

(1) Whereas in rubiginosus red forms a complete border around the
pileum (see frontispiece), it is “evanescent over the eye” in aerugi-
nosus ( Salvin and Godman 1892).

(2) The form aeruginosus is clear olive green on back, breast, and under-
parts, whereas rubiginosus is usually orange-olive in these areas.

(3) The form rubiginosus is banded below with narrow horizontal bands,
whereas the transverse bands on the breast and belly of aeruginosus
are wider and hastate or “V” shaped in pattern.

(4) The form aeruginosus is a larger bird than is rubiginosus (Blake
1953:290).

Similarities in aeruginosus and rubiginosus are (i) in color of pileum,
the latter being slate-gray in both ( Ridgway 1914, Blake 1953), and (ii ) fe-
males of both forms are identical in head coloration, i.e. both have slate-
gray pileums with red restricted to the nape region.

MENSUILVL CHAFL\CTERS

Males of aeruginosus average longer in bill length than rubiginosus [ =
P. r. yucatanensis) (Fig. 2), however, the difference is not statistically sig-

sononens! s

I 2

Q uriculoris

oerugi nosu s—n

f—

1 9

1 ^

oerug in osus—s

j 13

yucatanensis

max! mu s

168

THE WILSON lUJLLETIN • Vol. 90, No. 2, June 1978

TARSUS

5 onoriensis

a u riculoris

1 12

a e r u g I n osus-n

oer u g ! no su s-s
yucafonensis

m a xi m u s

1 —II

Fig. 4. Variation in tarsus length in Mexican Piculus. See Fig. 2 for explanation of
figure.

nificant. Females of both forms have very similar hill lengths. Both sexes of
aeruginosus are significantly larger in wing and tarsal length than ruhiginosus
(Figs. 3, 4). Although females of the 2 forms overlap in wing length, males
do not. Although males of the 2 forms overlap in tarsus length, females do
not. Males of northern aeruginosus ( from Nuevo Leon, Tamaulipas, and
San Luis Potosi ) average longer in wing length than southern aeruginosus
(from Puebla and Veracruz I , hut this difference is not statistically significant.

HEAD COLOR.ATION

Hargitt (1890:82) described a male aeruginosus from “Atoyac, Mexico”
with its crown color intermediate to the former and typical ruhiginosus in
having “red over the eye carried forward in a very narrow line to the base
of the bill.” Miller et al. (1957) refer to this specimen as having been taken
in Atoyac, Guerrero. Although a locality by that name does exist in Guerrero,
this would place it right in the middle of the range of P. auricularis. 1 think
this is highly unlikely since no field investigators subseijuent to Hargitt have
reported aeruginosus for that state, yet several auricularis have been taken
there. It is my belief that the Atoyac referred to is in fact near Orizaba in
Veracruz (see Gazetteer in Loetscher 1959:19).

Lawrence Binford (pers. comm.) independently came to the same con-

Baptista • REVISION OF MEXICAN PICULUS

169

elusion with different evidence. Alono; with P. aeruginosus, Binford found
that the collector (Mrs. Smith) also collected Celeus castaneus, Momotus
lessoni, Trogon puella, Rhamphastos carinatos, and other Atlantic species at
Atoyac, making it unlikely that the latter is in Guerrero. I have examined
Hargitt’s (loc. cit.) specimen in the British Museum of Natural History ( BM
98-3-10-130) and have since found 13 other individuals possessing crowns
with various amounts of red intermediate to “typical” aeruginosus and
rubiginosus. The following scoring system has been devised to describe
variation in crown color:

Score

1. Red nuchal patch continuing to just behind or immediately over the
eye (“typical” aeruginosus).

2. Red nuchal patch as in #1 but some small flecks of red in front of eye
or over lore.

3. Red nuchal patch as in #1 but thin red line from above eye to the base
of the bill [as in Hargitt’s (1890) specimen].

4. Red nuchal patch with pileum conspicuously surrounded by red
( “typical” rubiginosus ) .

Data on crown color are summarized in Table 1. Two of 3 males taken
48 km east of Huachinango, Puebla (9.5 km west from the Veracruz line)
had intermediate ( class 3 ) crowns. One male taken 4.75 km east of the road
to Villa Juarez, Puebla, had a class 2 crown. One male taken at Presidio,
35 km south of Cordoba, Veracruz had an intermediate ( class 3 j crown. In
all other characters it was typical of rubiginosus. Two other males taken
15.8 km away had crowns conspicuously surrounded with red; however,
comparison with material taken farther south indicated that these 2 Vera-
cruz males had less red surrounding the pileum than “typical” rubiginosus.
Three other males from Veracruz with class 3 crowns included single individ-
uals from Huatusco, Atoyac, and Tampico. Two males from Jalapa, Vera-
cruz had class 2 crowns.

In addition to the intermediates taken in central Veracruz or the near
vicinity, 5 other intermediates were examined from the more northern states.
These include a male from Linares, Nuevo Leon (ML 42758) (crown class
2j, 2 males from Rio Corona (ML 40289) and Ciudad Victoria, Tamaulipas
(BM 98-3-10-124) (crown class 2), a male from above Ciudad Victoria,
Tamaulipas (crown class 3) ( BM 9-8-3-10-123 ), and a male taken 47.5 km
east of Ciudad Maiz, San Luis Potosi (ML 32469) (crown class 2). A male
taken 5 km north of Gomez Farias, Tamaulipas (HC 4586) has a class 3

crown.

170

THE WILSON BULLETIN • VoL 90, No. 2, June 1978

Number of Specimens

Table 1

OF AERUGINOSUS AND VARIANTS

Color States

Showing

Different

Crown

State

Score
2 3

Nuevo Leon

Tamaulipas

San Luis Potosi

Puebla

Veracruz

* Series taken near Presidio, Veracmz.

These data indicate that birds with crown classes intermediate to aeruginosus
and rubiginosus may be found through the entire distributional range of
aeruginosus.

BODY COLOR

I found the color of the back to be a good index of general body coloration.
The following scoring system was devised to study variation in back color.

1. Clear olive green back (reference specimen ML $ 32473, Ciudad Maiz,
San Luis Potosi) .

2. Green back with a slight flush of yellow (reference specimen ML $
54405, 27 km east of Tapanatepec, Chiapas).

3. Green back with darker yellow flush than #2. Sometimes darker orange
tips to feathers of back (reference specimen ML $ 45499, Finca
Cacahuatl 24 km northeast of Tapanatepec, Chiapas).

4. Orange-green back, darker than #3 (reference specimen ML $ 35271,
Socoltenango, Chiapas).

Back classes 1 and 2 represent “typical” aeruginosus and back class 4
“typical” rubiginosus (= P. r. yucatanensis ) . Data on back scores are sum-
marized in Table 2. It may be seen that in the northern parts of its range,
aeruginosus tends to be lighter, whereas in the south it is darker.

A female aeruginosus from Papantla, Veracruz, had a back score of 3.
Another female from Poza Rica was given a back score of 3 but tended
towards a 2. A male from Puebla and a male from Linares, Nuevo Leon
also had back scores of 3. Three of a series of 6 rubiginosus taken at Presidio,
Veracruz and 3 birds from Tenozapa, Veracruz had back scores of 3. Thus,
in central Veracruz and nearby Puebla, individuals of the form aeruginosus
may be similar in back color to some individuals of rubiginosus. Moreover,

Baptism • REVISION OF MEXICAN PICULUS

171

Back

Table 2

Color in aeruginosus and yucatanensis

Score

State

Nuevo Leon

Tamaulipas

San Luis Potosi

Puebla

Veracruz- 1

Veracruz-2*

* “Pure” yucatonensis from contact areas.

the darker rubiginosus-Wke back (class 3j appeared as far north as Nuevo
Leon.

PATTERN OF BARRING ON BREAST AND BELLY

The proportion of hastate spotting to barring and the width of the bars
on breast, belly, and sides vary greatly in aeruginosus. The following 3 point
scoring system was devised to study pattern of barring on the undersides of
aeruginosus.

1. V or U shaped bands on entire breast, upper parts of belly, and sides
(reference specimen ML 9 42961, Linares, Nuevo Leon, see Fig. 5j.

2. V or L shaped bands restricted to upper breast. Lower breast and all or
almost all of belly and sides with horizontal bands. Yellow bands often
narrower than in #1 (reference specimen ML $ 32469, Ciudad Maiz,
San Luis Potosi) .

3. Horizontal bands on entire breast, belly, and sides. Yellow bands as
narrow or narrower than in #2 (reference specimen 5 51606, Huatusco,
Veracruz.

Barring class 3 represents the condition in “pure” rubiginosus (= P. r.
yucatanensis ) . Barring class 2 represents the intermediate form, and class 1
the “pure” aeruginosus. In class 1 the pattern on the underside gives one
the impression of yellow spots or scales. Sometimes the shapes of the bars
in breast types 1 and 2 are similar to those in Piculus auricularis, a fact
noted earlier by Salvin and Godman (1889). Bars in auricularis are hori-
zontal, each band divided in the center by a very small V. Banding scores
are summarized in Table 3.

A male taken at Huatusco, Veracruz (ML 51606) was light green on the
back (back score 2) with red stopping behind the eye, characters of aerugi-

172

THE WILSON BULLETIN • VoL 90, No. 2, June 1978

Fig. 5. This is a negative print to emphasize barring patterns on skins of yucatanensis,
aeruginosus, and variants. From left to right: 1. ML 9 34B09, typical aeruginosus,

taken 30 miles east of Huachinango, Puebla. 2. ML 3 40613, typical aeruginosus, taken
at Rio Corona. 18 miles north of Ciudad Victoria, Tamaulipas. 3. ML $ 32469, variant
aeruginosus with horizontal bars intermixed with hastate patterning, taken 30 miles east
of Ciudad Maiz. San Luis Potosi. 4. ML S 51606. variant aeruginosus completely
barred below as in yucatanensis, taken 9 miles south of Huatusco, Veracruz, Mexico. 5. ML
9 45492. typical maximus for comparison, taken 15 miles northeast of Tapanatepec,
Chiapas, Mexico.

nosus. It was, however, indistinguishable in banding pattern from pure
rubiginosus.

A female taken at Jalapa, Veracruz ( AM44069 ) was almost entirely barred
below with horizontal \ rubiginosus ) bands. Only a few very small L-shaped
spots on its upper breast indicate some aeruginosus ancestry. It was, how-
ever, aeruginosus green (back score 2) on its back. A similar female taken
at Cordoba, Veracruz, is in the collection of the British Museum of Natural
History ( BM 1857-7-30-4 I . It was taken with a male bearing the hastate mark-
ings of a “pure” aeruginosus ( BM 1857-7-30-5 j on the breast with the belly
by degrees coming close to rubiginosus-iy^e barring (Goodwin in litt.j. A
male taken at La Gloria, 15.8 km northwest of Presidio, Veracruz, (1 of a
series of 3 males and 1 female ) had a handing pattern reminiscent of aeru-
ginosus in being almost scale-like. The other 3 birds had typical rubiginosus
horizontal bands.

Baptista • REVISION OF MEXICAN PICL LUS

173

Banding

Table 3

Pattern in Breast and Belly

ON AERIGIXOSIS

State

Score

Nuevo Leon

Tamaulipas

^an Luis Potosi

Puebla

Veracruz

It is clear that individuals with banding patterns intermediate to aerugi-
nosus and rubiginosus (Table 3i may be found in all but one state. Nuevo
Leon. Moreover. 3 individuals had banding patterns indistinguishable from
’’pure'* rubiginosus.

Ridgway ( 1914 1 records both rubiginosus and aeruginosus from Mount
Orizaba. He also reports an aeruginosus from Cordoba. Veracruz; a variant
described above was taken 55.4 km from this town. Additional variants
taken at Cordoba in the collection of the British NIuseum have been de-
scribed earlier. Lowery ( 1951 1 reports taking an aeruginosus at Portrero
Viejo. 31.6 km from Presidio. Veracruz. I have described variant and "pure’"
rubiginosus taken 15.8 km from Presidio. Specimens showing characters
typical or intermediate to both forms in various combinations have been
found throughout the range of aeruginosus (Tables 1-3 1 . This suggests
that the latter has not differentiated completely from rubiginosus and must
be regarded as a race of this species following Miller et al. (1957 i.

VARIATION IN P. r. yucatanensis AND P. r. maximus
Griscom ( 1929 ( described the race maximus from Guatemala as a larger
bird than yucatanensis and greener throughout, lacking the "golden brown
wash” of the latter race. Miller et al. ( 1957 i included maximus in the Mexican
check-list restricting its range in Mexico to the Pacific slope of extreme
southeastern Oaxaca and adjacent Chiapas highlands at moderate altitudes.

P. r. maximus is separable from yucatanensis on the basis of wing length in
both sexes (Fig. 3i. ing length in male maximus averaged 130.8 mm and
that in yucatanensis averaged 119.2 mm i one-sided p < 0.00005 • . Mean
wing length in female maximus was 128.1 mm and that in female yucatanensis
was 117.0 mm (one-sided p < 0.00005 i .

} ucatanensis are highly variable in coloration. The same 4-point scoring
system used to study color variation in aeruginosus was used to study color
variation in yucatanensis and maximus (Table 4i. Back color was again

174

THE WILSON BULLETIN • Fo/. 90, No. 2, June 1978

Table 4

Back Color in yucatanensis and maximus

Color

Score

1-2

2-3

3-4

yucatanensis

maximus

taken as a rough index of general body color. A score of 4 indicated a
typical yucatanensis and a score of 3 or less a typical maximus.

The difference in color scores (Table 4) between my large-winged samples
and my small-winged samples is significant (p < 0.0001, Wilcoxon). How-
ever, since there is a great overlap in scores, color alone is not a good char-
acter to separate the races.

Presidio, Tenozapa, and La Gloria, Veracruz are in the vicinity of the zone
of contact between yucatanensis and aeruginosas. Six lighter birds from
these localities (scores 3 or tending towards 2, Table 4j may be indicative
of introgression of aeruginosas genes. Similarly, a light bird from Comitan
(1 of 4) may reflect some maximus genes. However, 5 lighter colored birds
from Catemaco, Matias Romero, and Tumbala (scores 2 to 3) are probably
illustrative of the erratic type of color variation in yucatanensis remarked on
by Griscom ( 1929 j .

One maximus from near Tapanatepec, Chiapas (scores 3^) and 2 dark
birds from Socoltenango (Table 4) may reflect some yucatanensis ancestry.
However, 1 dark bird from Cacahuatan and 1 from Pijijiapan (scores 3-4
to 4) are probably erratic variants as discussed above.

Howell (1952) noted that red pigment in sapsuckers iSphyrapicus) is
restricted to the tips of their feathers, a fact applying also to Piculus. Griscom
( 1929) found that the feathers of the back in males may sometimes be red-
tipped. I observed this occasionally, e.g. in a female from Catemaco, Vera-
cruz and a male from Las Palmitas, Honduras. Red may also invade the
top of the pileum in various degrees. These observations suggest that the
P. rubiginosus complex shared a common ancestor with Piculus rivoli; the
latter as a red-backed species variable in the amount of this pigment on the
pileum and back and with patterns below suggesting rubiginosus.

SIZE DIMORPHISM BETWEEN THE SEXES

Size dimorphism between the sexes was computed as percent difference in
mensural characters between males and females and coefficient of difference
( Mayr et al. 1953 ) .

Baptista • REVISION OF MEXICAN PICULUS

175

Table 5

Sexual Dimorphism in

PlCULUS

c.d.

% joint
nonoverlap

BILL

sonoriensis

15.4

1.3

auricularis

8.5

0.65

<75

aeruginosus-^

12.1

1.3

aeruginosus-S

13.2

1.5

yucatanensis

6.2

0.55

—

maximus

7.4

0.70

<75

WING

sonoriensis

2.3

0.67

auricularis

0.18

0.04

—

aeruginosus-^

5.6

1.5

aeruginosus-S

4.3

0.99

yucatanensis

1.8

0.47

—

maximus

2.1

0.38

—

TARSUS

sonoriensis

2.2

0.35

—

auricularis

5.5

0.38

—

aeruginosus-^

5.5

1.1

aeruginosus-S

4.2

0.47

—

yucatanensis

3.3

0.32

—

maximus

0.6

—

All samples are most dimorphic in bill length, as evidenced by percent
difference in measurements and percent of joint nonoverlap (Table 5j. Only
in yucatanensis is dimorphism in bill length not significant. There is also
a reduction in dimorphism from north to south, i.e from sonoriensis to
auricularis and from aeruginosus to yucatanensis. The difference in di-
morphism between northern and southern samples of aeruginosus is small
and probably due to sample error.

There is also a north to south reduction of dimorphism in wing length,
from sonoriensis to auricularis and from aeruginosus to yucatanensis. The
slight north to south reduction in dimorphism between the samples of aerugi-
nosus is due to the slight north to south decrease in wing length in male
aeruginosus.

Only northern aeruginosus are dimorphic in tarsus length. This again
may be due to sample error.

Davis (1965:566) has suggested that “the evolution of accentuated sexual

176

THK W ILSON BULLETIN • VoL 90, No. 2, June 1978

dimorphism in a given character is one means of increasing the variability
of the character in the population. But another method would be the evolution
of accentuated individual variability within each sex in the character con-
cerned, with presumed increased variability in diet.” Selander (1966) in-
terprets dimorphism, especially in bill length, as adaptive radiation at the
intra-populational level so that individuals may “occupy different subniches
or adaptive subzones, subdividing and, perhaps, expanding the total zone
or niche” used by the population. Kilham (1965), Selander (1966), and
others ( review in Hogstad 1976 I have studied a variety of woodpecker species
and found that either each sex fed on different species of trees, employed
different methods of foraging, or fed on different portions of the same tree.
However, Ligon (1968) studied Red-cockaded Woodpeckers iPicoides bore-
alis) which were only slightly sexually dimorphic for bill size and found dif-
ferences in foraging sites between the sexes. Subsequently, Ligon ( 1973 )
studied White-headed Woodpeckers [P. albolarvatus ) with greater dimor-
phism in bill length (10%) than any other measurement and found no dif-
ferences in foraging sites between the sexes. He cautioned, therefore,
( Ligon 1973:867) that one cannot always predict foraging patterns by de-
gree of sexual dimorphism in bill size. The Mexican forms of Piculus differ
greatly in degree of sexual dimorphism in bill size, from 15.4% in sonoriensis
or 13.2% in aeruginosas to almost none in yucatanensis. It is hoped that these
data will encourage others to conduct ecological studies on this little known
group.

DISCUSSION

In this treatment, all taxa of Piculus proposed in the Mexican check-list
(Miller et al. 1957) are recognized, with, however, differences in the ranges
of the subspecies of P. auricularis (Fig. 1). The western auricularis complex
has been described as paler than the rubiginosus forms of the east and south-
east. Buchanan ( 1964 ) has observed a similar situation for Least Pygmy
Owls [Glaucidium niinutissimum ) . This trend could probably also be found
in other Mexican birds with similar distribution ranges.

Size decreases clinally from north to south (Figs. 2-4), i.e. from sonorien-
sis to auricularis and from aeruginosas to yucatanensis, a manifestation of
Bergmann’s rule ( Mayr 1942, James 1970, Mengel and Jackson 1977). The
dines are steeper for males than they are for females for wing and bill length.

The tendency towards darker individuals in the southern parts of the ranges
of both the auriculnris and rubiginosus complexes is a manifestation of
Gloger’s rule ( Mayr 1942 ) . Gloger’s rule may also be interpreted as
the result of selection for crypticity. Selander and Giller (1963)
discuss the color patterns found in Melanerpes aurijrons, pointing out

Baptista • REVISION OF MEXICAN FICULUS

177

that “Boldly patterned types (M. a. aurijrons and M. a. polygrammus) ,
which are in gross aspect much lighter than the narrowly barred types [M. a.
dubius and M. a. santacruzi ) , are found in arid regions where light penetrates
deep into the middle and lower strata of relatively open woodland vegetation,
and the background of trunks and branches to which the woodpeckers are
exposed is relatively light in color. But in more humid regions where denser
broad-leaved vegetation creates greater areas of shadow and where the color
of the vegetation tends to be darker, the dark appearance of the narrowly
barred form is probably at a selective advantage. In similar fashion we
assume that the value and hue of the color of the breast and sides are adaptive,
providing effective counter-shading by being darker in more humid areas
and lighter in those of greater aridity.” Concommitant with color changes in
the Piculus forms mentioned are also changes in the vegetation types as shown
in Leopold’s (1959 ) vegetation map of Mexico. In Chiapas both Bergmann’s
and Gloger’s rules operate locally to produce the larger, lighter race maximus
of the montane districts of the Pacific Cordillera and the smaller darker race
yucatanensis of the Atlantic lowlands.

I have discussed the results of selection against red pigment on the crowns
of aeruginosus and auricularis. This perhaps may be interpreted as selection
for increased crypticity in areas of more open vegetation where bright colors
would render them more conspicuous. It is noteworthy that in none of the
forms is the red of the male malar stripes in any way affected. Noble (1936 )
has shown the importance of such malar stripes in sexual recognition in the
Common Flicker [Colaptes auratus). It is probably because of a similarly
important role as a social releaser that these stripes are retained in auricularis
and aeruginosus.

SUMMARY

Variation in the Mexican representatives of the genus Piculus was studied in order
to determine the status of the described forms (species and subspecies). A qualitative
analysis of museum specimens representing population samples was presented along
with quantitative data including statistical treatments of hill, wing, and tarsus measure-
ments.

The status of Piculus auricularis sonoriensis Van Rossem and Hachisuka was discussed:
the type and a topotype examined did not differ from samples taken in neighboring states
as indicated by the authors. However, northern samples of Piculus auricularis from
Sonora to Colima were separable from material from Guerrero and Oaxaca on the basis
of size and color. Two races are thus recognized.

Variation in Piculus aeruginosus Malherbe was studied and evidence was presented
suggesting genetic continuity and introgression with P. r. yucatanensis so that it is here
regarded as a race of rubiginosus. Two races of Piculus rubiginosus are recognized for
Chiapas, maximus of the Pacific G)rdillera and yucatanensis of the Atlantic lowlands,
separable on the basis of coloration and wing length.

All the forms of Piculus in the Mexican check-list (Miller et al. 1957) are thus recog-

178

THP: WILSON BULLETIN • VoL 90, No. 2, June 1978

nizcd with some changes in geographic distribution; these were divided into 2 species
complexes including the gray crowned auricularis forms occupying the more xeric west and
the slate crowned ruhiginosus races of the more mesic east and southeast. In aeruginosus
and auricularis size decreased clinally from north to south and intensity of coloration was
found to increase in a clinal fashion in the same direction in accordance with Gloger’s
rule. Manifestation of Gloger’s rule was discussed as being possibly the result of
selection for crypticity. Both laws seem to operate locally in Chiapas to produce the
larger, lighter colored highland race nuiximus.

Size dimorphism between the sexes was found to decrease from north to south in
both species complexes. Size dimorphism was most pronounced in bill length in all
but one form (yucatanensis) . The possible ecological significance of these findings was
discussed.

ACKNOWLEDGMENTS

My thanks first of all to Robert T. Orr who called my attention to this problem, and
for all his advice and assistance in the preparation of this paper; to Jacqueline Schone-
wald and my wife, Joyce, for their help and advice in the statistical methods; to Jerome
A. Jackson, Ned K. Johnson, Edward Kessel, Lloyd Kiff, David Mullen, Robert T. Orr, and
Lester L. Short for reading various versions of this manuscript and for their valuable
suggestions; to the late Elwood Molseed for providing transportation to Mexico City
where I examined some critical material and for his delightful companionship in the
field.

My thanks also to the following institutions and individuals for the loan of material
which made this study possible and for courtesies extended to me during my visits
to some of these institutions: The American Museum of Natural History^ (AM) (L. L.
Short), the Allan Phillips Collection at the Institute de Biologia in Mexico City (DM)
(A. R. Phillips, B. Villa, A. Villalobos), the Western Foundation of Vertebrate Zoology
(HC) ( E. Harrison, L. Kiff, and the late J. S. Rowley), the Dickey Collection at the
University of California, Los Angeles (UCLA) (G. Bartholomew, T. Howell, H. Pough
and C. Rischer), the Robert T. Moore Collection at Occidental College (ML) (J. W\
Hardy), Louisiana State University (LSU) (G. Lowery), the Museum of Zoology, Uni-
versity of Michigan (MU) (R. W. Storer), the Museum of Vertebrate Zoology, Univer-
sity of California, Berkeley (MVZ) (N. K. Johnson), the British Museum of Natural
Histor>’ (BM) (D. Goodwin), and the Delaware Natural History Museum (DM) (D.
M. Niles).

A grant from the California Academy of Sciences enabled me to do some field work
in the Sierra de Miahuatlan in Oaxaca in the summer of 1%5. In the summer of 1966
an additional grant from the California Academy of Sciences and the courtesy of Dr.
Elwood Molseed permitted more field work in Mexico and the visit to Dr. Allan Phillips'
collection. A grant from the American Association for the Advancement of Sciences
enabled me to visit the institutions in Southern California listed above. For making
these grants available to me, my sincerest thanks to Drs. Robert T. Orr and George E.
Lindsay.

Last but not least, I thank Dr. George M. Sutton for the beautiful colored plate which
adds much to this paper.

LITERATURE CITED

Baldwin, S. P., H. C. Oberiiolser, and G. Worley. 1931. Measurements of birds.
Scientific Publ., Cleveland Mus. Nat. Hist. 2:1-165.

Buptista • REVISION OF MEXICAN PICULUS

179

Belcher, C. and G. D. Smooker. 1936. Birds of the colony of Trinidad and Tobago.
Part IV., Ibis 6:792-813.

Blake, E. R. 1953. Birds of Mexico. Univ. of Chicago Press.

Buchanan, 0. M. 1964. The Mexican races of the Least Pygmy Owl. Condor 66:
103-109.

Davis, J. 1965. Natural history, variation, and distribution of the Strickland’s Wood-
pecker. Auk 82:537-590.

Dickey, D. R. and A. J. Van Rossem. 1938. The Birds of El Salvador. Field Mus. Nat.
Hist., Zool. Ser. 23:1-609.

Edwards, E. P and R. B. Lea. 1955. Birds of the Monserrate area, Chiapas, Mexico.
Condor 57:31-54.

AND R. E. Tashian. 1959. Avifauna of the Catemaco Basin of southern Vera-
cruz. Condor 61:325-337.

Gehlbach, F. R., D. O. Dillon, H. L. Harrell, S. E. Kennedy, and K. R. Wilson.
1976. Avifauna of the Rio Corona, Tamaulipas, Mexico: northeastern limit of
the tropics. Auk 93:53-65.

Griscom, L. 1929. Studies from the Dwight collection of Guatemalan birds I. Am.
Mus. Novitates 379:1-13.

Hardy, J. W. 1966. Physical and l)ehavioral factors in the sociality and evolution of
certain parrots ( Aratinga) . Auk 83:66-83.

. 1967. Rhynchopsitta terrisi is probably a valid species: a reassessment.

Condor 69:527-528.

Hargitt, E. 1890. Catalog of the Pieariae in the collection of the British Museum.

Catalog of the birds of the British Museum 18:1-597.

Hogstad, 0. 1976. Sexual dimorphism and divergence in winter foraging behavior
of three-toed woodpeekers Picoides tridactylus. Ibis 118:41-50.

Howell, T. R. 1952. Natural history and differentiation in the Yellow-bellied Sap-
sucker. Condor 54:237-282.

James, F. C. 1970. Geographic size variation in birds and its relationship to climate.
Ecology 51 :365-390.

Kilham, L. 1965. Difference in feeding l)ehavior of male and female Hairy Wood-
peckers. Wilson Bull. 77:124-145.

Leopold, A. S. 1959. Wildlife of Mexico. Univ. of Calif. Press, Berkeley and Los
Angeles, California.

Ligon, j. D. 1968. Sexual differences in foraging behavior in two species of
Dendrocopos woodpeckers. Auk 85:203-215.

. 1973. Foraging behavior of the White-headed Woodpecker in Idaho. Auk

90 :862-869.

Loetscher, F. W. 1955. North American migrants in the state of Veracruz, Mexico.
Auk 72:14-54.

Lowery, G. H. and W. W. Dalquest. 1951. Birds from the state of Veracruz, Mexico.

Univ. of Kansas Publ., Mus. of Nat. Hist. 3:531-649.

Martin, P. S., C. R. Robins, and W. B. Heed. 1954. Birds and hiogeography of the
Sierra de Tamaulipas, an isolated pine-oak habitat. Wilson Bull. 66:38-57.

Mayr, E. 1942. Systematics and origin of species. Columbia Univ. Press, New \ork.
, E. G. Linsley, and R. L. Usinger. 1953. Methods and principals of syste-
matic zoology. McGraw Hill Book Co., New York.

180

THE WILSON BULLETIN • Vol. 90, No. 2, June 1978

Mengkl, R. M. and J. a. Jackson. 1977. Geographic variation of the Red-cockaded
Woodpecker. Condor 79:349-355.

Miller, A. H., H. Friedmann, L. Griscom, and R. T. Moore (eds.). 1957. Distribu-

tional check-list of the l)irds of Mexico. Part II. Pac. Coast. Avif. 38.

Noble, G. K. 1936. Courtship and sexual selection of the flicker {Colaptes auratus
luteus). Auk 53:269-282.

Orr, R. T. 1966. Vertebrate biology. W. B. Saunders Co. Press, Philadelphia.
Peters, J. L. 1948. Check-list of birds of the world. Vol. 6. Harvard Univ. Press,
Cambridge.

Peterson, R. T. and E. L. Chalif. 1973, A field guide to Mexican birds. Houghton
Mifflin, Boston.

Ridgway, R. 1914. Birds of North and Middle America. U.S. Natl. Mus. Bull. 50
(6) : 1-882.

Robins, C. R. and W. B. Heed. 1951. Bird notes from La Joya de Salas, Tamaulipas.
Wilson Bull. 63:263-270.

Rowley, J. S. 1966. Breeding records of the Sierra Madre del Sur, Oaxaca, Mexico.

Proc. Western Foundation Vert. Zool. K3):l-204.

Salvin, 0. AND F. D. Godman. 1889. Notes on Mexican birds. Part 2. Ibis 1:380-382.

. 1892. Biologia Centrali Americana. Taylor and Francis, London.

ScHALDACH, W. J., Jr. 1963. The avifauna of Colima and adjacent Jalisco, Mexico.

Proc. Western Foundation of Vert. Zool. Kl):l-100.

Selander, R. K. 1966. Sexual dimorphism and differential niche utilization in birds.
Condor 68:113-151.

AND D. R. Giller. 1963. Species limits in the woodpecker genus Centurus

lAves). Bull. Am. Mus. Nat. Hist. 124:213-274.

Short, L. L. 1965. Hybridization in the flickers i Colaptes) of North America. Bull.
Am. Mus. Nat. Hist. 129:309-428.

— . 1972. Systematics and behavior of South American Flickers (Aves, Colaptes) .

Bull. Am. Mus. Nat. Hist. 149:1-109.

. 1973. The Green-barred Flicker and Golden Green Woodpecker of South

America. The Living Bird 12:51-54.

Skutch, a. F. 1943. The family life of Central American woodpeckers. Sci. Monthly
56:358-364.

. 1948. Life history of the Golden-naped Woodpecker (Tripsurus chrysauchen) .

Auk 65:225-260.

. 1956. Roosting and nesting of the Golden-olive Woodpecker P. r. uropygialis.

Wilson Bull. 68:118-128.

. 1969. Life histories of Central American birds HI. Pac. Coast Avif. 35.

Sutton, G. M. 1951. Mexican birds. Univ. Oklahoma Press, Norman.

. 1953. Bronzed Woodpecker. Wilson Bull. 65:65-67,

AND 0. S. Pettingill, Jr. 1942, Birds of the Gomez Farias region of south-
western Tamaulipas. Auk 59:1-34.

Van Rossem, A. J. 1945. A distributional survey of the birds of Sonora, Mexico.
Occas, Pap. Mus. Zool., Louisiana St. Univ. 21:1-379.

— AND The Marquess Hachisuka. 1937. A new woodpecker of the genus

Piculus from Sonora. Proc. Biol. Soc. Wash. 50:195-196.

Wetmore, a. 1941. Notes on the birds of the Guatemalan highlands. Proc. U.S.
Natl. Mus. 89:523-581.

Baptista • REVISION OF MEXICAN PICULUS

181

1943. The birds of southern Veracruz, Mexico. Proc. U.S. Natl. Mus. 93:215-

340.

. 1968. The birds of the Republic of Panama. Part 2. Columbidae (Pigeons)

to Picidae (Woodpeckers). Smithson. Misc. Coll. 150(2).

MOORE LABORATORY OF ZOOLOGY, OCCIDENTAL COLLEGE, LOS ANGELES, CA.
90041. ACCEPTED 20 DEC. 1977.

Wilson Bull., 90(2), 1978, pp. 182-196

DISTRIBUTION, DENSITY, AND PRODUCTIVITY OF
ACCIPITER HAWKS BREEDING IN OREGON

Richard T. Reynolds and Howard M, Wight^

Density of nests and productivity of Sharp-shinned Hawks {Accipiter
striatus). Cooper’s Hawks {A. cooperii) . and Goshawks (A. gentilis) within
Oregon are of interest because of recent declines of accipiter hawks in the east-
ern United States ( Schriver 1969, Hackman and Henny 1971, Henny and
Wight 1972). One factor implicated in this decline was contamination with
chlorinated hydrocarbons ( Ratcliffe 1970, Cade et al. 1971, Anderson and
Hickey 1972, Wiemeyer and Porter 1970). Snyder et al. (1973) presented
data on levels of DDE in eggs of accipiter hawks from various regions in North
America, including Oregon. Their data indicated that eggs of each species are
contaminated, but they were unable to evaluate the effects of contamination on
populations in Oregon as historical data on the abundance of breeding
accipiters did not exist.

This paper presents information on the distribution of nests, nesting density,
and nesting success of Sharp-shinned H-awks, Cooper’s Hawks, and Goshawks
in Oregon. In an attempt to assess current production trends, nesting densities
and productivities of Oregon accipiters are compared to densities and pro-
ductivities of accipiters elsewhere in North America and, where appropriate,
in Europe.

METHODS

This study included a survey for accipiter nests in all major forest types in Oregon
except the western juniper (Juniperus occidentalis) forests in central Oregon and the
Sitka spruce ( Picea sitchensis) forests along the northwest coast. Forests included in
this survey contain a wide variety of tree species, though with few exceptions, conifers
are dominant. These forests are primarily restricted to montane areas and vary from the
extensive and continuous forests of the Coast and Cascade ranges to the disjunct forests
of smaller mountain ranges east of the Cascades.

We divided Oregon into 3 subregions: (1) the Coast Range and the west slope

of the Cascade Range, a moist, densely forested region with a mild maritime climate,
referred to as western Oregon; (2) southwestern Oregon, which includes the Siskiyou
Mountains, characterized by relatively warm, wet winters and hot, dr>^ summers; and (3)
eastern Oregon (including the east slope of the Cascade Range), a high elevation and
more dry region with affinities to the Rocky Mountain forests. Franklin and Dyrness
( 1973) presented a list of the vegetational zones and associated tree species plus the
edaphic and climatic characteristics of these sub-regions. Nesting success and distribution
of nesting pairs of accipiters were determined during the breeding seasons of 1969 through
1974. During 1969 and 1970, all forest stands were searched for nests. However, after

^ Deceased.

182

Reynolds and Wight • ACCll’lTEK HAWKS IN OREGON

183

learning to recognize the structural characteristics of forest stands selected hy each
species for nesting, the amount of forest intensively searched was reduced. Searching
only stands considered potential nest sites proved suitable for surveying large areas,
hut undoubtedly caused us to overlook some nests. .Searches conducted during the 6
breeding seasons disclosed 117 nest sites and 1.39 attempts at nesting.

Nest density for each species was determined by intensively searching the Corvallis and
Rly study areas. The Corvallis study area, 9284 ha, was located on the east slope of
the Coast Range approximately 8 km northwest of Corvallis, Benton Co. (T. lOS and
11.^^; R. .3W and 6W) fFig. 1). Elevation of this area ranged from 80 to .300 m and was
of moderate relief. Except for some narrow valley bottoms Henced pastures) and one
burn in various stages of regeneration, the forests of this study area were continuous.
Douglas-fir (Pseudotsuga menziesii) was the dominant tree species and it existed in
pure stands or mixed with western hemlock (Tsuga heterophylla) , red alder (Alnus
rubra), or higleaf maple (Acer macrophyUum) . On some dry, south-facing slopes,
small stands of f)regon white oak (Quercus garryana) persisted. A major portion of
the Corvallis study area was composed of young ( C^lOO years) stands of Douglas-fir,
though stands of all age classes were represented. This area was searched during the
nesting seasons of 1970 and 1971.

The Bly study area, 11,741 ha. was in the Gearhart Mountains approximately 24 km
northeast of Bly, Lake Co. i T. 1.5E and 16E; R. 36.^^) < Fig. 2). Elevation of this area
ranged from 14.30 m to 21.30 m and was also of moderate relief. Except for 2 burns which
were in young regenerative stages and small natural openings, forests of this area were
continuous. Tree species composition varied from pure stands of ponderosa pine (Pinus
ponderosa) at lower elevations (southwest portion), through mixed stands of ponderosa
pine and white fir (Abies concolor) at mid-elevations, to mixed and pure stands of white
fir and lodgepole pine i Pinus contorta) at high elevations (north and east portions).
.Stands of all age classes in each timber type were represented; however, the most common
type was mature ponderosa pine overstory with mixed understory of ponderosa pine and
white fir. The Bly area was searched during the nesting season of 1974.

Data from these 2 areas also provided mean distances between nests of conspecifics.
.Since several researchers (e.g.. Hoglund 1964) reported only distances between nests, we
include a mean distance to make the dispersion of nests in our study areas comparable.
We determined this hy locating nests on maps and measuring the distance between each
active nest and its nearest neighbor, using the distance between any 2 nests only
once, .“^ince some pairs used different nest sites from year to year, distances between
nests were calculated on a yearly basis, using each nest as a single observation. Nest
sites were visited up to 2 months after fledging to determine the length of time young
remained in the nest area under care of the adults.

RESULTS AND DISCUSSION

Distribution of nesting pairs. — Of the 139 nesting attempts, 16 were Sharp-
shinned Hawk, 42 were Cooper’s Hawk, and 81 were Goshawk.

All 3 species were found nesting in eastern, western, and southwestern
Oregon. Sharp-shinned Hawks nested in the Coast and Cascade ranges, the
Siskiyou, Gearhart, and the Steens mountains at elevations ranging from 120
m in the Coast Range to 2010 m in the Gearhart Mountains. Cooper’s Hawks
nested in the Coast and Cascade ranges, the Siskiyou, Wallowa, and Gear-

THE WILSON BULLETIN • Vol. 90, No. 2, June 1978

Fig. 1. Corvallis study area. Illustrating main drainages and accipiter nest locations
in 1969, 1970, and 1971.

hart mountains. This species also nested in the floor of the Willamette Valley
( western Oregon ) in isolated, but extensive stands of Douglas-fir. Elevations
of Cooper’s Hawk nests ranged from 15 m in the Willamette Valley to 1760 m
in the Gearhart Mountains. Goshawks nested on both east and west slopes
of the Cascade Range, the Siskiyou Mountains, and in all mountain ranges
in eastern Oregon. Elevation of Goshawk nests ranged from 580 m on the
west slope of the Cascades to 1860 m in the Gearhart Mountains. Nests of
Goshawks were not found in the Coast Range.

The relative abundances of nests of each accipiter species in our statewide
sample possibly reflected the relative difficulty of locating nests rather than
their actual abundances. Goshawks, because of their large size, are the

Reynolds and Wight • ACCIPITER HAWKS IN ORECiON

185

10 12

• I — I

Fig. 2. Bly study area. Illustrating main drainages and accipiter nest loeations in 1974.

easiest to locate; whereas Sharp-shinned Hawks, the smallest accipiter in
Oregon, are the most difficult. This bias in our sample makes it appear that
Sharp-shinned Hawks are the least abundant of the Oregon accipiters, while
the opposite may be true (see below). We feel that, with the possible excep-
tion of Sharp-shinned Hawks in Western Oregon, the relative density of each
species in the Corvallis and Bly areas approximates the actual densities of
nesting accipiters in each region.

Nest densities and spacing. — Portions of the Corvallis study area received
a cursory nest search in 1969; 2 Cooper’s and 1 Sharp-shinned hawk nests
were located. Four Cooper’s Hawk nests were located in 1970 and 5 in
1971 (Fig. 1), resulting in 1 nest per 2321 ha in 1970 and 1 nest per 1857 ha
in 1971. Mean distance between nests was 5.0 km in 1970 ( range = 3. 7-6.3
km, SD = 1.29 km) and 5.5 km in 1971 (range = 4.S-6.9 km, SD = .97
km). No nests of Sharp-shinned Hawks were located in this study area in
1970 or 1971.

186

THE WJLSON BULLETIN • Vol. 90, No. 2, June 1978

Table 1

Nest Site Tenacity of Accipiter Hawks in Oregon, 1969-1974

Number of Years

of Occupancy

R/R*^

Reoc-

cupancy

Rate

Numlier of nest

sites revisited

Sharp-shinned

Number of nest

Hawk

sites reoccupied

3“

2/11

.18

% reoccupancy

Number of nest

sites revisited

Cooper’s Hawk

Number of nest

sites reoccupied

16“

7/34

.21

% reoccupancy

Number of nest

sites revisited

Goshawk

Number of nest

sites reoccupied

20“

25/63

.40

% reoccupancy

“ Number of sites occupied for 1 year only.

Number of sites reoccupied/number of sites revisited.

In 1974, 4 Sharp-shinned Hawk nests, 5 Cooper’s Hawk nests, and 4
Goshawk nests were located in the Bly study area resulting in an overall
density of 1 nest per 903 ha or 1 Sharp-shinned Hawk nest per 2750 ha, 1
Cooper’s Hawk nest per 2200 ha, and 1 Goshawk nest per 2750 ha (Fig. 2).
Mean distances between nests of conspecifics were: Sharp-shinned Hawks,
4.1 km (range = 1. 8-6.0 km, SD = 2.12 km) ; Cooper’s Hawks, 3.5 km (range
= 2.0-4.4 km, SD = .79 km) ; and Goshawks 5.6 km (range = 2.4-8.4 km,
SD = 3.00 km ) (Table 1 ) .

Two Sharp-shinned Hawk nests were approximately 300 m from active
Cooper’s Hawk nests, and 1 Sharp-shinned Hawk nest was approximately
450 m from an active Goshawk nest. Five Cooper’s Hawk nests were between
300 to 450 m from active Goshawk nests. In 2 of the latter cases, both species
used the same nest sites for 2 consecutive years. In all the above situations,
only 1 nest, a Sharp-shinned Hawk adjacent to a Cooper’s Hawk, failed to
fledge young.

Mean distance between nesting pairs of Cooper’s Hawks in the Bly and
Corvallis areas was considerably greater than the approximate 1.6 km
lietween nests found in Arizona ( N. Snyder pers. comm.) and in California
(Fitch et al. 1946). Meng (1951) did not determine a mean distance between

Reynolds and W ight • ACCIPITER HAWKS IN OREGON

187

30 nests of Cooper’s Hawks in New York, but reported that the 2 closest
nests, on opposite edges of 2 extensive woods separated by a large field, were
2.4 km apart.

Mean distance between adjacent pairs of Goshawks in the Ely area was
essentially the same as reported for European Goshawks [A. g. gentilis) in
Sweden (5.5 km, range = 3.9-8.0 kmj (Hoglund 1964 j. While McGowan
(1975) did not report a mean distance between any of 9 active nests in
Alaska, he found a density of 1 pair per 4869 ha in 1971 and 4142 ha in 1972,
only half the density of Goshawks in Oregon. In Finland, a density greater
than that in Oregon has been reported — 1629 ha per pair ( 9 pairs) (Hakila
1968).

Nest site tenacity. — Nest sites of accipiters were defined as the portion of a
forest stand containing the nest and the requisite structural features of the
vegetation (density, height, canopy closure) and physiographic conditions
(slope, aspect, surface water) used by a nesting pair during the breeding
season. In Oregon, the area within a nest site increased with increasing
accipiter size, ranging from approximately 4 ha for Sharp-shinned Hawks, 6
ha for Cooper’s Hawks and 8-10 ha for Goshawks.

Many established nest sites of each accipiter contained more than 1 nest.
We did not observe Sharp-shinned or Cooper’s hawks reoccupying an old nest.
If a pair of either species returned to a previously used site, a new nest,
usually within 100 m of the old, was constructed. In contrast, many pairs
of Goshawks used the same nest for 2 or more years or alternated between 2
or more nests within an established site. Alternate nests within an established
Goshawk site varied from 15 to 150 m apart, though most were 60-90 m
apart. In addition, several pairs of Goshawks had alternate nest sites, usually
within 0.4 km, between which they shifted on a 1-3 year basis. In general,
nest site tenacity increased with increasing accipiter size. The maximum
number of years a nest site was occupied was 2 years for Sharp-shinned
Hawks, 3 years for Cooper’s Hawks, and 5 years for Goshawks (Table 1).

For one reason or another an established pair of Goshawks may desert
one nest site for another up to 3.5 km away. For example, in 1974 one pair
of Goshawks on the Ely study area, which used one site for several years,
moved 3.2 km north to a site which was approximately 2.4 km south of another
active Goshawk nest. Since these hawks were not marked, we were not certain
that both nest sites had been occupied by the same pair. However, frequent
sightings of the male foraging within areas used in previous years and carrying
prey from these areas toward the new site suggested that the same pair
was involved. Nest site shifts of this type occasionally placed conspecific
pairs in close proximity (less than 3.5 km apart), though none of these
situations was found to persist for more than one breeding season.

188

THE ILSON BULLETIN • \ ol. 90, \o. 2, June 1978

Three shifts of nest sites, averaging 3.1 km (range = 2. 6-3.2 kmj were
also noted for Cooper's Hawks in the Corvallis study area. Although not
certain the same pairs were involved, we observed both single and pairs of
hawks flying between old and new nest sites on several occasions before egg
laying.

In 2 instances, nest sites of 1 species were occupied during subsequent breed-
ing seasons by other species. The first of these involved a site used for 2
seasons by Sharp-shinned Hawks prior to its being used in the third year
by Cooper's Hawks. Another involved a site used by Goshawks for 2 years
prior to its being occupied by Cooper's Hawks. No interactions between pairs
involved were noted.

Initiation of breeding and egg laying. — One pair of Goshawks was first
noted in its nest site on 23 March, and most pairs were found in their sites
by early April. In Oregon, the date of clutch completion and initiation of
incubation by Goshawks was highly variable. The earliest clutch was com-
pleted about 10 April, and the latest, 2 June iFig. 3). However, most
clutches were completed and incubation began within the last week of April
and the first 2 weeks of May, approximately the same period reported by
McGowan I 1975 I for Goshawks in interior Alaska.

Regression analysis suggested little association between the date of initiation
of incubation and the elevation of 30 Goshawk nests in Oregon, for all years
combined as well as in any single year.

One pair of Cooper's Hawks was first noted in its nest site on 28 March
I western Oregon ) , while most pairs throughout Oregon were observed in or
about the nest sites by mid-April. In western Oregon, clutches were completed
and incubation began during the last week of April through the third week of
May, while those in eastern Oregon were completed during the third through
the last week of May (Fig. 3j. The earliest completed clutch for Cooper's
Hawks was 1 May i western Oregon ), and the latest, 30 May i eastern Oregon).

One pair of Sharp-shinned Hawks was first noted in its nest site on 9 May,
6 days before the first egg was laid. On this date the nest was complete, in-
dicating that the hawks had probably been at the site for at least a week prior
to 9 May. Clutches of Sharp-shinned Hawks were completed and incubation
began sometime in May, although some may not be completed until mid-June
( Fig. 3j . The earliest completed clutch was 14 May and the latest 19 June.

Due to the narrow range of elevation over which we found nests within
each subregion and because our sample of Cooper's and Sharp-shinned hawk
nests is not continuous over the entire range of elevation from western to
eastern Oregon (highest in western Oregon, 600 m; lowest in eastern Oregon.
1400 m). it was difficult to determine whether or not there was an association
between elevation and date of nesting for these species. However, initiation

ACCIPITER HAWK- IN OREGON

189

Reynolds and •

AES~E-S

EiS'E=S

AES'

26 ‘.'AY

A = = :L ‘.<AY .JSE

Fig. 3. Number • :f nesis and approximate date initiation of incubation in Sharp-
'bime-i Hawk. G>*p^r*' Hawk, and Gfhawk in Or^on. 1969-1974. i:-uthwe~tem
Or^n nc ~ho-v*Ti due to in?uffi>: lent numiiYrr of neats.

of incubation in both species in eastern Oregon was somewhat later than in
western Oregon -Fig. 3'. In general, clutch size decreased as the size of
accipiter increased Table 2 • .

Incubation and nestling period. — In Oregon, the incubation period lasted
3*>-32 days for each accipiter species. This was about the same period re-
porte»i by McGowan • 1975 • for Goshawks in Alaska 29 days '. Bent ■ 1937 '
listed '28 days for Goshawks. 24 days for Cooper's Hawks, and 21-24 days for
^aarp-shinned Hawks. Brown and Amadon 1968'. who summarized much
of the literature concerning birds of prey, reported incubation periods of 36
days for Cooper's Hawks and 34—35 days for Sharp-shinned Hawks.

]90

THE WILSON BULLETIN • Vol. 90, No. 2, June 1978

Table 2

Mean Number of Eggs Laid, Eggs Hatched, and Young Fledged per Nest for Sharp-
shinned Hawks, Cooper’s Hawks, and Goshawks in Oregon

Species

Year

Mean number
of eggs

Mean number of
eggs hatched

Mean number of
yoimg fledged/
nest attempt

Sharp-shinned

1969

5.0 (3)“

3.7 (3)“

2.3 (3)"

Hawk

1970

—

1971

3.0 (1)

1.0 (1)

.5 (2)

1972

5.0 (1)

4.0 (Ij

4.5 (2)

1973

—

4.0 (1)

1974

—

3.0 (3)

Total

4.6 (5j (0.89)’’

3.2 (5) (1.30)”

2.7(111 (1.74)”

Cooper’s Hawk

1969

4.3 (4)

4.0 (4)

2.0 (3)

1970

4.0 (1)

3.3 (3)

1971

3.8 (4)

2.0 (4)

1.6 (7)

1972

3.5 (4)

2.3 (4)

2.0 (7)

1973

—

2.5 (2)

1974

—

2.0 (2)

Total

3.8(13) (1.14)

2.8(13) (1.91)

2.1(24) (1.56)

Goshawk

1969

—

1.5 (2)

1970

3.5 (2)

3.0 (2)

1.8 (4)

1971

3.0 (2)

2.0 (2)

1.6 (5)

1972

3.0 (1)

1.9(22)

1973

—

1.5(11)

1974

—

2.0 (4)

Total

3.2 (5) (0.45)

2.6 (5) (0.89)

1.7 (48) (0.76)

“ Number of nests.

** Standard deviation.

Hatching of all eggs in Sharp-shinned and Cooper’s hawk clutches occurred
in 1 or 2 days. The time required for hatching of all eggs in a Goshawk
clutch was not determined. Nests of each species frequently contained 1 and
sometimes 2 eggs that did not hatch. Goshawks covered these with short,
green fir houghs, hut in Sharp-shinned and Cooper’s hawk nests, these eggs
frequently remained exposed. Unhatched eggs eventually broke and the shell
fragments disappeared.

The nestling period lasted 34-37 days for Goshawks, 27-30 days for
Cooper’s Hawks, and 21-24 days for Sharp-shinned Hawks. Faster develop-
ment of the smaller males and their subsequent earlier fledging accounted
for much of the variation in length of nestling period within each species.

Mean numlier of Sharp-shinned Hawks fledged in Oregon was below that

Reynolds and Wight • ACCIPITER HAWKS IN OREGON

191

Table 3

Mean Clutch Size and Number of Young Fledged per Nest of Sharp-shinned Hawk,
Cooper’s Hawk, and Goshawk in this Study Compared to Others

Species

Source

Location

Year

Clutch

size

Number

fledged/

nest

attempt

Sharp-shinned

This study

Oregon

See Ta-

4.6 (5)“

2.7

(11)“

Hawk

Craighead and

ble 2

Craighead, 1956

Wyoming

1947

3.5 (2)

3.5

(2)

Cooper’s Hawk

This study

Oregon

See Ta-

3.8(13)

2.1

(24)

Craighead and

ble 2

Craighead, 1956
Craighead and

Michigan

1942

4.3 (6)

2.0

(6)

Craighead, 1956

Michigan

1948

4.0

2.3

(7)

Meng, 1951

New York

1948-50

4.2(36)

—

Henny and

1929-45

3.53’’

(118)

Wight, 1972

Northeastern

1949-67

2.67*=

(54)

U.S.

Goshawk

This study

Oregon

See Ta-

3.2 (5)

1.7

(48)

ble 2

McGowan, 1975

Alaska

1971

3.1(10)

2.5

(10)

McGowan, 1975

Alaska

1972

3.0(14)

1.8

(14)

McGowan, 1975

Alaska

1973

3.8 (9)

1.8

(9)

Hakila, 1968

Finland

1955-58

3.4(22)

1.5

(28)

Holstein, 1942

Denmark

1937-40

2.8 (9)

1.8

(9)

Hoglund, 1964

Fennoscandia

1931-40

3.1(11)

“ Number of nests.

** Number of young reaching bandable age per successful nest 1929—1945.
' Number of young reaching bandable age per successful nest 1949—1967.

reported for the same species in Wyoming in 1947 (Craighead and Craig-
head 1956; Table 3). Number of Cooper’s Hawks fledged per nest in Oregon
was slightly above the number fledged in Michigan in 1942 and slightly
below the number fledged in the same area in 1948 (Craighead and Craig-
head 1956; Table 3). The number of Cooper’s Hawk young fledged per
successful nest in Oregon was considerably less than the number of young
reaching bandable age per successful nest in the years prior to the introduc-
tion of organochlorine pesticides and slightly above the number in later
years in northeastern United States — 2.9 per successful nest in Oregon com-
pared to 3.53 in 1929-1945 and 2.67 in 1949-1967 (Henny and Wight 1972;
Table 3 j . However, since mortality between the time of handing and fledging
was not determined, Henny and Wight’s (1972) figures were overestimates
of productivity and not directly comparable to our data.

192

THE WILSON BULLETIN • Vol. 90, No. 2, June 1978

Table 4

Hatching Success/ Fledging Success/ and Percent of Total Nests of Sharp-shinned
Hawks, Cooper's Hawks, and Goshawks in Oregon that Fledged at Least One Young

Species

% Hatching
Success

% Fledging
Success

% Successful
Nests

Sharp-shinned

Hawk

69.9 ( 5 nests )“
(23 eggs)^

81.2 ( 5 nests)
(16 young)

91.7(12 nests)*

Cooper’s Hawk

74.0 ( 13 nests)
(50 eggs)

61.4(14 nests)
(44 young)

69.0(29 nests)

Goshawk

81.2 ( 5 nests)
( 16 eggs)

72.0 (11 nests)
(25 young)

90.4(52 nests)

1 Number of eggs hatched /number of eggs laid.

2 Number of yormg fledged/number of young hatched.
“ Total number of nests.

Total number of eggs laid.

Total number of young hatched.

Number of Goshawks fledged per nest in Oregon was nearly the same
as reported in southwestern Finland ( Hakila 1968), Denmark (Holstein
1942), and near Fairbanks, Alaska in 1972 (McGowan 1975), but was
nearly 1 young less per nest than in Alaska during 1971 (McGowan 1975;
Table 3). However, mean clutch size was nearly the same for all of these
locations. Similarities in clutch size in Oregon and Alaska and the relatively
high fledging success in Alaska in 1971 suggested that clutch size for this
species may be somewhat constant geographically and annually, while hatch-
ing and fledging success are influenced by food availability. For example,
snowshoe hares {Lepus americanus ) were the primary food of Goshawks in
Alaska during 1970-72 (McGowan 1975). In northern latitudes hare popula-
tions are subject to an approximate 10-year population cycle (Keith 1963).
In the area of Fairbanks, Alaska hare populations peaked in 1971 and de-
creased in numbers in 1972 through at least 1973 (McGowan 1975; J.
Ernest pers. comm.). High hare density probably accounted for the high
fledging success ( 2.5 young per nest ) in 1971. In 1972, however, the number
fledged per nest decreased to 1.8 young, and was again 1.8 in 1973 (McGowan
1975). In addition, the number of nests occupied in McGowan’s study area
was 7 in 1971, 9 in 1972, 8 in 1973, and 1 in 1974.

Age of nesting hawks. — In the North American accipiters, both sexes begin
molting into adult plumage in the spring of their first year. Since this molt
is not completed until the following fall, nesting accipiters can be identified
as immature or adult ( 2 or more years old ) on the basis of plumage. Of 70
Goshawk females and 10 Sharp-shinned Hawk females we observed nesting,
all were in mature plumage. However, 2 females (6%) of 34 pairs of

Reynolds and Wight • ACCIPITER HAWKS IN OREGON

193

Cooper’s Hawks were immature. Males were not observed at all of these
nests, but of those seen of each species, all were in mature plumage. Meng
(1951) in New York, and N. Snyder (pers. comm.) in Arizona found several
immature female Cooper’s Hawks, but no immature males breeding. McGowan
(1975) found 4 of 11 Goshawk females nesting in Alaska in immature
plumage in 1971, while in 1972 and 1973, all were in mature plumage. He
observed males at 37% of the nests, and each of these was mature. Hoglund
(1964) reported that immature female Goshawks were occasionally found
nesting in Finland. On the basis of an examination of testes of 10 immature
male Goshawks, all of which varied in size and only 1 of which contained
small amounts of mature sperm, Hoglund (1964) concluded that immature
males are normally incapable of breeding. However, Glutz von Blotheim
(1971) reported that 2 of 30 male (6.7%) and 9 of 93 female (9.7%)
Goshawks nesting in central and southern Europe were in immature plumage.

Reynolds ( 1972 ) discussed the general lack of nesting by immature males
and hypothesized that, since males are the principal food providers during the
nesting season, foraging experience is a prerequisite for successful nesting.
Immature males, lacking experience, may be subject to greater risks of pre-
dation or accident while foraging, and may spend greater energies in ter-
ritorial establishment and defense than mature birds. Deferring the age of
first breeding should increase the future fitness of an immature male. A con-
comitant of deferred maturity is delayed testicular growth and spermatogenesis.

Post-fledging period. — Young accipiter hawks, as the young of most birds
of prey, are dependent on adults for food for some time after fledging. During
the early portion of this period fledged young remain close to the nest. As
flying skills develop, attachment wanes and young can be found at increasing
distances, though their activity remains centered around the nest. Decreased
attachment during the latter portion of this period increases the probability of
overlooking fledged young when visiting a nest site. For this reason, it is
difficult to determine exactly when parental care is terminated, and we report
the greatest number of days (the potential length of dependency) young were
found in or near the nest site after fledging.

Young of Sharp-shinned Hawks are the most inconspicuous of the 3 species
and most difficult to follow once fledged. Eight days was the longest period
we observed Sharp-shinned Hawks in the nest site following fledging. How-
ever, a pair in Utah, kept under surveillance by radio-telemetry, remained in
the nest area for nearly a month before leaving as a family group (Platt
1973). In Oregon fledgling dependency may persist as long as 42 days for
Goshawks and 53 days for Cooper’s Hawks. We concur with Ashmole and
Tovar (19681 who hypothesized that extended fledgling periods in predaceous

194

THE WILSON BULLETIN • Vol. 90, No. 2, June 1978

birds occurs in species whose prey are difficult to capture, as considerable time
is required for development of necessary hunting skills.

Current production trends. — To attempt a relative appraisal of production
of accipiters in Oregon, we have included clutch sizes and fledging rates of
accipiters from other populations (Table 3). Comparing these production
figures to those from Oregon showed that, except for mean clutch size in
Cooper’s Hawks and mean number of Sharp-shinned Hawks fledged per nest
(discussed below), clutch size and fledging rates in Oregon are either greater
than or are within the range of the figures from other populations. Of the 3
accipiters in Oregon, Sharp-shinned Hawks had the highest percentage of
successful nests (91.7), although Goshawks were close behind (90.4). How-
ever, the percent of successful nests of Cooper’s Hawks was considerably
lower (69.0) (Table 4), with 75% of the failures occurring in western Oregon.
Causes of nest failure in Cooper’s Hawks ranged from predation upon
nestlings (avian, 1 case; mammalian, 1 case), nest destruction (logging, 1
case; windstorm, 1 case), desertion of eggs (1 case) and an unexplained
occurrence in which the young either died or were killed and subsequently
eaten by the adults (2 cases). Causes of failure of 2 additional Cooper’s
Hawk, 1 Sharp-shinned Hawk, and 2 Goshawk nests were unknown, while
human disturbance caused nest desertion by the same pair of Goshawks during
2 consecutive years.

Hatching success (number of eggs hatched/number of eggs laid) was lowest
for Sharp-shinned Hawks (69.6%), intermediate for Cooper’s Hawks (74.0%),
and highest for Goshawks (81.2%), while fledging success (number of young
fledged/number of young hatched) was highest for Sharp-shinned Hawks
(81.2%), intermediate for Goshawks (72.0%), and lowest for Cooper’s
Hawks (61.4%) (Table 4).

Since all nests of Sharp-shinned Hawks hatched young (the 1 unsuccessful
nest failed after hatching) , the low hatching success of this species resulted pri-
marily from a high incidence of egg loss within clutches through infertility,
death of embryo, and egg breakage. Each of these factors, especially egg break-
age, may be related to the very high levels of pesticides in eggs reported for
Sharp-shinned Hawks in Oregon (Snyder et al. 1973).

Although egg loss is an important factor reducing productivity of Cooper’s
Hawks, this species suffers its greatest losses during the nestling period.
Predation and the unexplained deaths of the young and subsequent consump-
tion by the adults are the 2 most important factors reducing fledging success.
Incidence of predation upon nestlings is higher for Cooper’s Hawks than either
of the other accipiters in Oregon and may be related to habitat used for
nesting. Nest sites of Cooper’s Hawks, particularly in western Oregon, are
sufficiently open to allo\N the entry of large, winged predators, e.g., crows

Reynolds and Wight • ACCII^ITER HAWKS IN OREGON

195

[Corvus brachyrhynchos) , ravens (C. corax), and Great Horned Owls {Bubo
virginianus) , while body size of Cooper’s Hawks may not be sufficiently large
to repel these predators.

Reproductive effort of Goshawks seems to he met with relatively high suc-
cess in all 3 elements presented in Table 4. Factors decreasing the hatching
success of this species are an occasional infertile or addled egg, while most
losses of nestlings occur within 10 days of fledging. During this period, dead
young were frequently found below nests. Causes of death could not he de-
termined. A few Goshawk eggs were collected in Oregon and analyzed for
pesticides. All had relatively low levels, a fact that is consistent with the
lack of any noticeable population decline in any region of North America
(Snyder et al. 1973) .

SUMMARY

Distribution of nests and nesting success were determined for Sharp-shinned Hawks
{Accipiter striutus) , Cooper’s Hawks (A. cooperii), and Goshawks (A. gentilis) in west-
ern, southwestern, and eastern Oregon during 1969 through 1974. Nesting density was
determined by intensively searching a 9284 ha area in western Oregon and an 11,741 ha
area in eastern Oregon. These searches produced 4 Cooper’s Hawk nests (1 nest/2321
ha) in 1970 and 5 nests in 1971 (1 nest/1857 ha) in western Oregon and 4 Sharp-
shinned Hawk nests (1 nest/2750 ha), 5 Cooper’s Hawk nests (1 nest/2200 ha), and 4
Goshawk nests (1 nest/2750 ha) in eastern Oregon. An analysis of hatching success,
fledging success, and number of nests that were successful showed that Sharp-shinned
Hawks suffered the greatest losses during the incubation period, and Cooper’s Hawks and
Goshawks during the nestling period.

ACKNOWLEDGMENTS

Collection of data on nesting accipiter hawks requires an enormous amount of man-
hours. Our expectations could not have been fulfilled without the help of personnel from
the U.S. Forest Service and Weyerhaeuser Company, especially G. Cornett, B. Heckel, T.
Bryan, D. Williams, R. Anderson, C. Anderson, K. Horn, B. Anderson, J. Hoppe, and
J. Jakabosky. Special thanks to E. Forsman, who helped with the field work from
the beginning to the completion of this study. Others who helped include W. Pike, J.
Tabor, and G. Lind.

The following reviewed and gave helpful comments on the manuscript: E. C. Meslow, J.
Wiens, L. Hunt, J. Hunt, J. Denton, and J. Crawford. This study was conducted and sup-
ported under the auspices of the Oregon Cooperative Wildlife Research Unit: Oregon
Department of Fish and Wildlife, Oregon State University, U.S. Fish and Wildlife Seiwice.
and Wildlife Management Institute cooperating. This is Oregon Agriculture Experiment
Station Technical Paper 4137.

LITERATURE CITED

Anderson, D. W. and J. J. Hickey. 1972. Eggshell changes in certain North American
birds. Proc. 15tb Int. Ornithol. Congr. :514-540.

Ashmole, N. P., and H. Tovar S. 1%8. Prolonged parental care in Royal Terns and
other birds. Auk 85:90-100.

196

THE WILSON BULLETIN • Vol. 90, No. 2, June 1978

Bknt, a. C, 1937. Life histories of Nortli American birds of prey. Part 1. U.S. Natl.
Mus. Bull. 167.

Brown, L., and U. Amadon. 1968. Eagles, hawks, and falcons of the world. Part 1
and 2. McGraw-Hill, N.Y.

Cade, T. J., J. L. Linger, C. M. White, U. G. Roseneau, and L. G. Swartz. 1971. DDE
residues and eggshell changes in Alaskan falcons and hawks. Science 172:955-957.

Craighead, J. J.. and F. C. Craighead, Jr. 1956. Hawks, owls and wildlife. Stackpole
Co., Harrisburg, Pa.

Fitch, H. S., B. Glading, and V. House. 1946. Observations on Cooper Hawk nesting
and predation. Calif. Fish Game 32:144-154.

Franklin, J. F., and C. T. Dyrness. 1973. Natural vegetation of Oregon and Washing-
ton. U.S. Forest Service, Pacific N.W. Forest and Range Exp. Station, Res. Pap.
PNW-8.

Glutz von Blotheim, N. 1971. Handbuch der Vogel Mitteleuropas. Vol. 4. Fal-
coniformes. Akademische Verlagsgesellschaft, Frankfurt am Main.

Hackman, C. D., and C. J. Henny. 1971. Hawk migration over White Marsh, Maryland.
Chesapeake Sci. 12:137-141.

Hakila, R. 1968. Kanahaukan (Accipiter gentilis) pesimabiologiasta satakunnassa.
Eripainos Porin Lintutieteellinen Yhdistys ry:n vousikirjasta 1968:52-54.

Henny, C. J., and H. M. Wight. 1972. Population ecology and environmental pollution:
Red-tailed and Cooper’s hawks. Pp. 229-250, in Population Ecology of Migratory
Birds: A Symposium. U.S. Dept, of Interior Wildlife Research Report 2.

Hoglund, N. 1964. Der habicht Accipiter gentilis Linne in Fennoskandia. Viltrevy 2:
195-270.

Holstein, V. 1942. Duehogen Astur gentilis dubius (Sparrman). H. Hirschsprung
Verlag. Kopenhagen.

Keith, L. B. 1963. Wildlife’s ten-year cycle. Univ. Wisconsin Press, Madison.

McGowan, J. D. 1975. Distribution, density and productivity of Goshawks in interior
Alaska. Alaska Dept, of Fish and Game. P-R Proj. Rep., W-17-445.

Meng, H. K. 1951. The Cooper’s Hawk Accipiter cooperii (Bonaparte). Ph.D. thesis,
Cornell Univ., Ithaca, N.Y.

Platt, J. 1973. Habitat and time utilization of nesting Sharp-shinned Hawks — a
telemetry study. Conference on Raptor Conservation Techniques. Colorado State
University. March 22-25, 1973. Typescript.

Ratcliffe, D. a. 1970. Changes attributable to pesticides in egg breakage frequency
and eggshell thickness in some British birds. J. Appl. Ecol. 7:67-115.

Reynolds, R. T. 1972. Sexual dimorphism in accipiter hawks: A new hypothesis.
Condor 74:191-197.

ScHRiVER, E. C., Jr. 1969. The status of Cooper’s hawks in western Pennsylvania. Pp.
356-459, in Peregrine falcon populations: their biology and decline (J. J. Hickey,
ed.). Univ. Wisconsin Press, Madison.

Snyder, N. F. R., H. A. Snyder, J. L. Linger, and R. T. Reynolds. 1973. Organochlo-
rines, heavy metals, and the hiologv of North American accipiters. BioSci. 23:300-
305.

V lEMEYER, S. N., AND R. D. PoRTER. 1970. DDE thins eggshells of captive American
kestrels. Nature 227:737-738.

DEPT. OF FISHERIES AND WILDLIFE, OREGON STATE UNIV., CORVALLIS 97331. AC-
CEPTED 2 FEB. 1977.

Wilson Bull., 90(2), 1978, pp. 197-214

SOCIAL AND FORAGING BEHAVIOR OF WARBLERS
WINTERING IN PUERTO RICAN COASTAL SCRUB

William Post

The foraging behavior and social relationships of the warblers (Parulidae)
during the breeding season have been studied extensively by Morse ( 1967a,
1968, 1971, 1973 j. Parnell (1969j examined the foraging behavior of mi-
grating warblers. Other than the studies of Eaton (1953) and Lack and Lack
(1972) no work has aimed specifically at investigating the ecology of war-
blers wintering in the tropics. My objectives in the present study were to de-
scribe and quantify the social and foraging behavior of parulids wintering in
Puerto Rican coastal scrub and where possible to compare their behavior with
that reported from other regions. The Puerto Rican study sites were chosen
because of the structural simplicity of the vegetation, and the relatively high
density of warblers. In addition, the lack of significant predators prompted
me to examine the question of what influence this might have on flocking be-
havior.

STUDY AREA AND METHODS

I conducted the study on El Guayacan Island (35 ha) and on La Cueva Island (20 ha)
on the SW coast of Puerto Rico, 2 km SW of La Parguera. These “islands” are penin-
sular, connected to the mainland by a narrow mangrove forest no more than 100 m wide.
Before 1960 both islands were grazed by goats. In 1962 the Puerto Rican government
leased the islands to the Caribbean Primate Research Center, which uses them as sites
for free-ranging rhesus macaque {Macaca mulatta) colonies. High populations of
monkeys have been on the island since 1966, and foraging by monkeys has kept the
vegetation stunted. Except for a few scattered trees, all vegetation is less than 3.5 m.
In addition, the islands are in a severe dry zone. Rainfall is only 35 cm per year
(12 yr average for La Cueva Island). Scrubby vegetation predominates: corcbo
(Pisonia albida), oxhorn bucida (Bucida buceras) , and gumbo-limbo (Bursera sim-
aruba) are the dominant trees. Shrubs or small trees are pigeon-berry {Bourreria
succulenta), Bumelia spp., Lantana involucrata, dildo i Cephalocereus royenii) , and
pricklypear {Opuntia rubescens) . This community type is described in detail by Glea-
son and Cook (1926) as the xerophytic forest of the Ponce limestone. The scrub zone is
bordered by blaek mangrove (Avicennia nitida) and red mangrove i Rhizophora mangle).

I observed warblers between 20 December and 15 April of 1975-76. Observations were
made between dawn and 09:30. I walked along narrow paths (“slow walk” of Lack and
Lack 1972) and, upon encountering a warbler, noted (1) its foraging site, e.g. whether
it was on a broad-leaved or narrow-leaved tree; (2) its height; (3) its position in the
vertical vegetation strata, e.g. canopy or subcanopy; (4) its position in the horizontal
vegetational strata — inner, middle, or outer. These parameters were recorded only once
for each individual. Then, for the same bird I recorded 5 consecutive foraging tactics,
e.g. leaf-gleaning, flycatching, or hovering. During the time the bird was in view, I

197

198

THK WILSON BULLETIN • VoL 90, No. 2, June 1978

also noted whether it engaged in hostile interactions, gave contact calls, and the nature
of its social affiliation (alone, part of a cohesive flock, or part of a stationary flock).
Total observation time was 47 h.

I conducted censuses at the same time that I made observations of foraging and so-
cial behavior. On different days than those on which I made observations, I operated
mist nets (12 m long; 30 mm mesh) from dawn to 09:30, to correspond to census times.
Five sites, with 5 nets at each, were used throughout the study period. Total net-hours
were 153.

RESULTS AND DISCUSSION

Population composition. — Censuses and mist-netting gave the same esti-
mates of population composition (Table 1). Only uncommon species such as
Yellovv-rumped Warbler and Bahama Yellow throat w ere missed by either one
or the other method. Differences may be explained by the fact that mist nets
only cover a space from ground to 2 m. The census method is probably more
accurate for conspicuous, usually canopy-feeding species such as the Cape
May and Prairie warblers. The absence of Adelaide’s Warbler from the mist-
net sample is explained by the fact that it was territorial, and no nets were
placed within its home range. By both methods, the most common species
on the study sites was the Cape May Warbler, followed by the permanent
resident Yellow Warbler.

Although Lack and Lack (1972: Table 3) do not give comparative census
and netting times for their Jamaica study sites, a comparison of the propor-
tions of species seen and captured in Puerto Rico with proportions seen and
captured in Jamaican lowland arid habitat shows differences. Overall they
saw^ 34 and captured 55 individual w^arblers, and the numerically dominant
species was Prairie Warbler (26% of those seen and 33% of those netted),
followed by Ovenbird (15%, 13%), Common YellowThroat (12%, 0), N. Par-
ula (9%, 7%), and Palm (9%, 0). Interestingly, the ground feeding Palm
Warbler and Common Yellow throat w ere not represented in their net sample.
The Cape May, although wintering in Jamaica, was not recorded in lowland
arid habitat there.

The results of Lack and Lack’s (1972: Table 4) censuses for all 9 lowland
dry limestone forest localities in Jamaica reveal a greater number of species
than for my 2 Puerto Rican sites: 19 species vs. 11, although the number
of individuals recorded per 10 h was almost the same: 131 in Jamaica and 129
in Puerto Rico. The total number of species of North American warblers
wintering in Jamaica and Puerto Rico is the same (18; Bond 1956), and the
higher number of species recorded by the Lacks is presumably due to the
greater number of sites they visited. My study areas are probably most simi-
lar to the Lacks’ arid cut-over habitat, where they saw only 16 individuals of
7 warbler species per 10 h (Lack and Lack 1972: Table 5).

Post • WARBLERS IN PUERTO RICO

199

Number of Warblers Seen and Mist

Table 1

-NETTED IN CoASTAL SOUTHWESTERN PUERTO RiCO

El Guayacan and La Cueva
study sites

Mainland opposite to
study sites

Species

Number seen
per 10 h

Number captured
per 100 net h

Number seen per 10 party
hours in dry foresU

Black-and-white Warbler
(Mniotilta voria)

4.8 (3.7)"

3.9 (3.5)"

0.3 (0.4)"

Prothonotary Warbler
(Protonotaria citrea)

1.3 (1.0)

2.6 (2.3)

0.1 (0.1)

N. Parula

(Parula americana)

17.6(13.7)

16.4(14.8)

12.1(16.5)

Yellow Warbler

{Dendroica petechia)

18.9(14.7)

19.6(17.7)

26.2(35.8)

Cape May Warbler
{D. tigrina)

41.0(31.9)

26.2(23.7)

2.1 (2.9)

Black-throated Blue
Warbler

( D. caerulescens)

0.1 (0.1)

Yellow-rumped W'arbler
i D. coTonata)

1.9 (1.5)

3.2 (4.4)

Yellow-throated Warbler
{D. dominica)

0.2 (0.3)

Adelaide’s Warbler
( D. adelaidae)

3.5 (2.7)

7.2 (9.8)

Blackpoll Warbler
(Z). striata)

0.4 (0.6)

Prairie Warbler
( D. discolor)

16.0(12.4)

11.8(10.7)

2.6 (3.6)

Palm Warbler
( D. palmarum)

2.0 (1.8)

3.9 (5.3)

Ovenbird

{Seiurus aurocapilliis)

1.3 (1.2)

0.1 (0.1)

N. Waterthrush
(S. novehoracensis)

12.7 (9.9)

11.8(10.7)

9.7(13.3)

Louisiana Waterthrush
(S. motacilla)

0.2 (0.3)

Bahaman Yellowthroat
{Geothlypis rostrata)

1.0 (0.8)

1.3 (1.8)

Hooded Warbler
( Wilsonia citrina)

0.7 (0.6)

0.1 (0.1)

American Redstart
iSetophaga ruticilla)

9.9 (7.7)

14.4(13.0)

3.4 (4.6)

Total

128.7(100.0)"

110.7(100.0)"

73.2(100.0)

^ Calculated from three Christmas bird counts, 1972-1974.

^ Percentages in parentheses.

® No difference between census and mist-net estimates of the relative numbers of the 8 species
that were recorded by both methods ( X“ = 9.6, d.f. = 7; 0.25 > P > .1 ).

200

THE WILSON BULLETIN • VoL 90, No. 2, June 1978

Table 2

Social Affiliations of Individual Warblers Observed in Puerto Rican Coastal Scrub

Species

Bird

alone

Individuals
in stationary
flock^

Individuals
in cohesive
flock!

Total

individuals
in flocks-

Black-and-white Warbler

Protlionotary Warbler

N. Parula

Yellow Warbler

Cape May Warbler

145

Prairie Warbler

N. Waterthrush

American Redstart

Total

150

220

340

^ All flocks were composed of more than 1 species.

2 Includes birds that were not classified as to whether they were in stationary or cohesive flocks.

In comparison to the Christmas bird count censuses of warblers on the ad-
jacent Puerto Rican mainland (Table 1), the results obtained on my study
sites differed mainly in the higher proportion of Cape May and Prairie
Warblers and lower proportion of Yellow Warblers that I recorded. Although
data are lacking, these differences are presumably due to vegetational dif-
ferences between the islands and adjacent mainland.

Social behavior. — Warblers wintering on the study sites were often mem-
bers of flocks (Table 2), but these flocks were stationary aggregations, usual-
ly organized around concentrations of insects. Such flocks correspond to
the “collections” of Lack and Lack (1972). The behavior of the Puerto Rican
flocks was similar to that described for the Jamaican flocks: birds were
often within a few meters of each other but moved about independently. The
Lacks’ stationary flocks had up to 8 individuals and usually no more than 2
of 1 species. I found stationary flocks of up to 25 individuals and some spe-
cies, notably Cape May and Yellow warblers, were represented by up to 7
individuals.

Few warblers, only 7.7% (31 out of 401: Table 2) were organized into
cohesive flocks, i.e., flocks that maintained their integrity as they moved
through the scrub. Such cohesive flocks were easily identified because their
members consistently followed each other, rarely remaining at one position
long.

Morse (1970) defined a flock as 2 or more birds in a group, formation of
which depended upon positive responses by these individuals towards one an-
other. Groups that form due to common responses of individuals to an ex-
trinsic factor such as localized water or food Morse termed aggregations.

Post • WARBLERS IN PUERTO RICO

201

Several workers have studied stationary flocks or aggregations that gather
at fruiting trees. For example, Leek (1972) described the behavior of war-
blers and other species aggregating at Cecropia trees in Puerto Rico. In the
tropics warblers and other species often gather to forage on insects flushed
by army ant swarms ( Willis 1966 a,b). Flocking behavior may be viewed as
an adaptation to enhance foraging efficiency ( Cody 1971, Krebs et al. 1972),
a means of improved protection against predators ( Moynihan 1962, Powell
1974), and a means of reducing intraspecific aggression (Barash 1974). The
relative importance of these factors probably varies with species and habitat.
In this study I was particularly interested in flocking behavior because of the
few aerial predators in the study areas. Willis (1973 ) correlated a widespread
absence of cohesive flocks in Puerto Rico with paucity of accipiter species.
The one locality where Willis found cohesive flocks, Maricao, also has Sharp-
shinned Hawks {Accipiter striatus) .

At the La Parguera study sites, I found 2 species of avian predators, Amer-
ican Kestrels (Falco sparverius) and Short-eared Owls {Asia flammeus) . A
pair of Kestrels lived on La Cueva in 1972-1973. In 1974 they moved to
El Guayacan, probably in response to new feeding habitat created by the
clearing of land. Direct observations suggested that this pair specialized on
house mice {Mus musculus) and rats {Rattus norvegicus) that were abun-
dant around monkey enclosures. Although the Kestrels flew over groups of
warblers I was watching, the warblers did not appear to respond. A single
Short-eared Owl appeared sporadically on El Guayacan, and it may occasion-
ally have taken small birds, although I saw it foraging only at twilight around
the monkey enclosures. On 19 February 1975 I flushed it from some dense
grass, and it perched on an exposed stub, whereupon it was mobbed by a
Black-and-white, a Northern Parula, a Cape May Warbler, and a Northern
Waterthrush, all of which approached within 5 m of the owl. After about 3
min, the warblers resumed foraging within view of the perched owl. Cats,
mongooses { Herpestes javanicus) ^ and monkeys, all common on the study
sites, did not seem to influence the behavior of foraging warblers. Twice I
saw a mongoose move under foraging Northern Parulas and American Red-
starts, none of which altered their behavior. Cats were active during crepuscu-
lar periods; rhesus monkeys seemed to ignore free-flying birds.

Sweep samples made at different warbler foraging sites (Table 3) re-
vealed that insects were locally concentrated. Shrubs such as Bumelia sup-
ported abundant populations of homopterans, while neighboring patches
of vegetation such as Lantana had few insects. Similarly, black mangrove
stands were infested with homopterans, while adjacent red mangroves were rel-
atively insect free. Such a patchy distribution of insects facilitated concentra-
tions of warblers. Homopterans were so abundant that birds foraging on Bu-

202

THE W ILSON BULLETIN • VoL 90, No. 2, June 1978

‘‘Out of 1786 homopterans, 1740 were 1 species (Petrusa epilepsis).
•' 570 were P. epilepsis.

Post • WARBLERS IN PUERTO RICO

203

melia often flushed 5-10 insects with each move. I commonly observed flocks
of up to 25 warblers foraging in a dense area of Bumelia about 100 m-. Al-
though the study was conducted during the dry season ( December-April I,
and the coastal scrub habitat has the superficial appearance of being barren,
insects such as Homoptera, which can puncture plant tissue to reach phloem
tissue, are abundant. Janzen ( 1973 ) commented upon the disproportionate
abundance of Homoptera on Caribbean islands, and he considered the num-
bers of Homoptera that he collected on Icacos Island, Puerto Rico (vegeta-
tionally similar to the La Parguera study sites), to be “phenomenal.” Other
than birds I saw few insect predators on the study sites. Ameiva and Anolis
lizards were uncommon on the study sites, although common on the adjacent
mainland, and their numbers are probably reduced by cats and mongooses. I
captured relatively few spiders in my sweep net samples (Table 3).

All warbler species were found most often in stationary flocks, and several
were found only in stationary flocks or alone ( T ellow arbler, American
Redstart, and A. aterthrush. Table 2). The American Redstart and Northern
Waterthrush, because of their foraging tactics (see below), may be predis-
posed to forage in one position for long periods, making them even more
likely members of stationary flocks. Northern Waterthrushes occupied win-
ter territories, as probably did Yellow Warblers, but both species apparently
left them to visit insect concentrations. From the point of view of motiva-
tion, being alone or in a stationary flock may be the same, since these flocks
are apparently passive assemblages. The determinants of social affiliation in
these assemblages thus appeared to be food distribution and the foraging
tactics of the birds.

As suggested by Cody (1971), cohesive flocking may be adaptive under
conditions of low food availability. Kepler and Kepler (1970) observed that
mountainous areas of Puerto Rico have fewer bird species than the
lowlands and related this to the reduced number of flying insects in the
mountains, perhaps due to heavy rainfall. It is significant that the only area
where illis (1973) found cohesive flocks was in the mountains, around
Maricao. Willis commented upon the low productivity of Maricao, relating
it to the area’s poor soil characteristics. He suggested that under the condi-
tions of low food availability there, birds may have large home ranges,
facilitating the formation of cohesive flocks. Morse (1970) demonstrated
that flock-joining species showed the greatest tendency to group when over-
all population density was lowest, and he viewed cohesive flocking as an
adaptation to improve foraging. Conversely, Morse ( 1967b j found that when
food was abundant. Brown-headed Nuthatches {Sitta pusilla) dropped out of
mixed-species flocks (of which they were usual members) to forage alone.

A possible further explanation for the lack of cohesive flocking in the

204

THE W ILSON BULLETIN • Vol. 90, No. 2, June 1978

warblers I studied is the lack of nuclear species. Moynihan (1962 ) ob-
served that migrants or winter visitors are joiners or followers of mixed flocks
of permanent residents such as tanagers or honey creepers, which by their
gregarious and conspicuous behavior act as focal points of flock integration.
On my study sites no residents acted as nuclear species. Puerto Rican Todies
{Todus mexicanus), although noisy, tended to remain in one area for long
periods. Bananaquits {Coereba flaveola) concentrated their activities around
flowers. Stolid Flycatchers \Myiarchus stolidus) foraged near groups of
warblers, but their foraging method probably precluded their membership in
cohesive flocks. The greater proportion of migrants which compose the avi-
fauna of islands, e.g. 44% for Puerto Rico vs. 26% for the Canal Zone (Leek
1972) may influence the frequency of cohesive flocking, due to the relative
rarity of potential nuclear species on islands.

Calling behavior. — I made 216 observations of the frequency that warblers
gave contact calls, the sibilant tseep or tsip notes often heard during migra-
tion. On 99 occasions (45.8%) warblers uttered contact calls (Table 4). The
occurrence of these calls was related to the social situation of the warblers.
Birds in flocks gave the calls significantly more often than did those alone
[X~ — 7.3; P < .01), and birds in cohesive flocks called more often than those
in stationary flocks (j^- = 19.6; P<.005). However, birds alone and in
stationary flocks called with the same frequency (X“ ~ This lat-

ter result is further evidence of the similarity between foraging in stationary
flocks and solitary foraging.

Although the data are few, some species differences in calling behavior are
evident (Table 4). Cape May Warblers called proportionally more often than
other species, whether alone or in flocks. At least in flocks, this could be due
to the numerical dominance of this species: contact calls may be responded
to more often by conspecifics. In contrast to other species, Prairie arblers
seldom called, whether in or out of flocks.

Aggressive interactions. — I recorded hostile interactions between warblers,
which included chases, supplants, and fights. As found by Morse (,1970),
most aggressive encounters were intraspecific (Table 5). This is due to
the fact that a bird is more likely to encounter a member of its own than
another species in its preferred foraging space. In addition, members of
the same species may be attracted by each others’ morphological and be-
havioral characteristics (Moynihan 1962). For example, on several occasions,
I saw dull-plumaged Cape Mays, presumably juveniles, following adult male
Cape Mays as they foraged. arblers that captured a large food item such
as a caterpillar were often chased by conspecifics.

If we assume that each species has an equal probability of encountering
another, then we may calculate the expected number of hostile interactions

Post • WARBLERS IN PUERTO RICO

205

206

THE WILSON BULLETIN • Vcl. 90, No. 2, June 1978

Table 5

Hostile Interactions of Warblers Wintering in Puerto Rican Coastal Scrub

Species attacked

Species attacking

Prothonotary

Warbler

N. Parula

Cape May Warbler

Prairie Warbler

N. Waterthrush

American Redstart

Unidentified Warbler

Prothonotary

Warbler

(14.1; 12.7-15.5)"
N. Parula

-(0)^

3(1)

— (1)

1(3)

1(1)

-d)

— (1)

(7.5 ±0.3)“
Yellow Warbler

—(0)

—(2)

2(2)

6(4)

— (1)

1(1)

(11.3 ±0.5)

Cape May Warbler

-(1)

13(9)

—(10)

44(22)

2(8)

—(6)

—(5)

(10.1 ±0.2)
Prairie Warbler

—(0)

1(1)

—(1)

-(2)

4(1)

— (1)

—(0)

(6.9 ±0.2)

N. Watei thrush

(15.1 ±0.6)
American Redstart

(7.4 ±0.3)

Unidentified

Warbler

Total

1 Mean weight ( g ) followed by range ( N = 5 ) .

2 Mean weight (g) followed by 95% confidence interval (sample size >10) for species other
than Prothonotary Warbler.

^Expected number of aggressive encounters, correcting for relative abundance (Table 1) and
assuming all species are equally likely to encounter each otlier (for species which gave 5 or more
attacks to identified species).

per species pair (Table 5: values in parentheses). A comparison of observed
and expected values shows that birds did attack members of their own species
more often than expected. The Cape May, which comprised 32% of the war-
bler population (Table 1) gave 62% of the attacks (Table 5), and 73% of
these were to conspecifics. Species which were similar in foraging behavior
were also attacked more frequently than expected; for example Cape Mays
vs. Parulas. The 1 species for which I have sufficient data may be ordered
into a linear hierarchy with Yellow Warhlers dominating Cape Mays, fob

Table 6

Foraging Heights of Warblers in Puerto Rican Coastal Scrub

Post • WARBLERS IN PUERTO RICO

207

c M

w >

QJ'O

it; c

O s

•r I)

”£ £

A c
c^5 §

p « £

’£ <u o

6'-

«
3 «

vd cd

O eo

id cd
m

I I

-r

0\ CM

CO 00

1-4 cm'

CO CM
cd MO

•$. -z ^

c w i,

^ 2-^

’c3 ^

cm'

,— 1

.— 1

1—^

'—1

1— 1

>-H

'—1

,—1

,— 1

cm'

c^'

,— ^

1— 1

111

208

THE WILSON BULLETIN • Vol. 90, No. 2, June 1978

'sJ a

5 ^

X ^

T) o>

K M

c u
CO ^

r-5 t-

0\ CO CO

CO i-H csi co‘

o O O fO lO

lO O CO o' (N O

(M ,— I ^

r- CO
LO 'C?

.-H o p

lo CO ON

o (M p C<j

r-4 CO o

CO o lO

CNICsILO OOONrHlO
(Mi-HOiOi— ICOCO-^
t— I C\|

Post • WARBLERS IN PUERTO RICO

209

Table 8

Structural Units of Vegetation Used for Foraging by Warblers in Puerto Rican

Coastal Scrub

Species

Number of
observations

Percentage of observations in:

Canopy

Subcanopy

Inner

Middle

Outer

Inner

Middle

Outer

N. Parula

1.7

5.0

31.7

15.0

13.3

33.3

Yellow Warbler

4.1

25.0

41.7

33.3

Cape May Warbler

103

1.0

5.8

40.8

9.7

16.5

26.2

Prairie Warbler

8.3

25.0

14.6

37.5

American Redstart

64.5

12.9

22.6

lowed by Northern Parulas, then Prairies. This is also the order of decreas-
ing weight (Table 5).

Foraging behavior. — My observations indicate that warblers wintering in
Puerto Rican coastal scrub often used similar foraging spaces and feeding
tactics (Table 6-9). Two species, the Black-and-white Warbler, a trunk and
branch gleaner, and the Northern Waterthrush, a ground feeder, are clearly
separated from the other species. Of the others, all of which concentrated
their foraging activities in above-ground foliage, the American Redstart is a
specialist of the inner subcanopy (Table 8), and it obtained most of its prey
on the wing (Table 9). The remaining 4 species, the Northern Parula, Yel-
low, Cape May, and Prairie warblers were not clearly separated in their
foraging behavior: The similarities among these species may be quantified
by using the index of overlap (Table 10) developed by Horn (1966) :

„ _ (Xi + Yi) log (Xi + Ti) - S X. log Xi - S Ti log

"" (A + T)iog(A + y)-AiogZ-yiogy

where X and Y equal the number of observations in samples of foraging of
the two species being compared; Xi ^nd y-i equal the proportion of observa-
tions in the foraging category of the samples.

The data show broad overlap in many categories of foraging behavior,
particularly in foraging site use and in foraging tactics. Less overlap was
shown among use of structural units. Northern Parulas were very similar
to Cape Mays in the 3 parameters considered. Parulas were also somewliat
close to Prairies but quite dissimilar to American Redstarts. The Cape May
overlapped broadly in at least 2 categories with each of the other 4 species
and, after the N. Parula, was most similar to the Prairie. The Yellow Warbler
was closest to the Prairie Warbler.

F()UA(;in(; 'I'aciics of Wauhi.khs in Hican Coastal Scuuh

210

THE WILSON BULLETIN • Vol. 90, .Vo. 2, June 1978

' Slriicliirtis less Ilian 10 inin in dianieler classilietl as Ivvij^s.

A liird en^afied in eliasinn wlien it llnllei<‘d or ran alonj' a liori/.ontal snrlaee in pnrsnil ol prey.
Wel>-leedin« liirds readied onl to take spiders or spi<lers' prey Ironi well or thread.

212

THK WILSON BULLETIN • VoL 90, No. 2, June 1978

Lack and Lack (1972 ) presented data on the foraging behavior of warblers
wintering in Jamaica. Unfortunately, they lumped data from different hab-
itats and localities, so it is not possible to quantify the overlap of feeding
behavior with confidence. They stated that most of their flocks were col-
lections, i.e., stationary flocks, perhaps attracted to locally abundant food,
and this implies that within habitats there could have been much overlap
among species. For example, 2 species in Jamaica that were not separated
by habitat. Cape May and N. Parula, were said to be separated by feeding
sites and foraging heights (Lack and Lack 1973: Table 11). These are
roughly equivalent to my foraging sites (Table 7) and foraging heights
(Table 6). The foraging site overlap value for these 2 species in Jamaica
w as 0.98 vs. my value of 0.97. In Jamaica the mean foraging height for these
2 did not differ significantly (N. Parula :6.4 m, 95% confidence interval ±0.6
m; Cape May : 7.6 ± 1.1 m). Comparable heights in Puerto Rico were N.
Parula :2. 5 ± 0.3 m; Cape May:2.8±0.3 m. In Jamaica, Prairie Warblers
and N. Parula Warblers were also not separated by habitat, and their feeding
site overlap is 0.99 (vs. 0.94 in Puerto Rico), while their feeding heights are
not significantly different (N. Parula :6.4 ± 0.6 m; Prairie :5. 3 ± 0.8 m vs.
Puerto Rican figures of 2.5 ± 0.3 and 2.1 ± 0.3). My data indicate that there
is less overlap among species in the structurally simple Puerto Rican habitat
than in Jamaica. However, the Jamaican data are from several habitats and
sites and thus may show more overlap due to the lumping of different data
sets.

The patches of food visited by the stationary flocks that I studied were
composed mainly of 1 species of insect (Table 3). Morse (1970) noted a con-
vergence in the foraging behavior and use of feeding site of Browm-headed
Nuthatches and Pine Warblers (Dendroica pinus) as they exploited heavy
crops of pine seeds, even though this resulted in increased numbers of ag-
gressive encounters.

When food is locally abundant and accessible, there may be little pressure
for feeding specialization by members of mixed flocks. Food may be ob-
tained by the simplest method, and if birds are using the same prey, con-
vergence in their foraging patterns w^ould be expected. North American
warblers wintering in localities with high food concentrations appear to show
few behavioral differences when foraging. As Willis (1966a) pointed out,
superabundance of food may he frequent rather than exceptional due to the
“irregularity principle”: available food is often left by irregularities in time
or space, since resource exploitation lags behind its appearance. In the
tropics North American migrants, because of their mobility and opportunistic
behavior, may often concentrate on irregularly distributed food that is not ef-
fectively used by resident populations.

Post • WARBLERS IN PUERTO RICO

213

SUMMARY

The population density of warblers wintering in arid coastal scrub in Puerto Rico was
much higher than comparable sites in Jamaica. Warblers wintering in the coastal scrub
usually foraged alone or in stationary (passive) flocks. Few were organized into co-
hesive (integrated) flocks. Warblers in stationary flocks gave contact calls with the same
frequency as those alone, while warblers in cohesive flocks called more often. Aggres-
sive behavior was common, and occurred most often between members of the same species
or those with similar foraging behavior. The stationary flocks were usually composed of
birds that gathered to harvest concentrations of insects whose distribution varied be-
tween vegetation patches. The few predators in the study sites apparently did not in-
fluence the flocking behavior of the warblers. Warblers showed considerable overlap in
their feeding behavior, which may be due to convergence of foraging on the same abun-
dant food source.

ACKNOWLEDGMENTS

I thank the Caribbean Primate Research Center for allowing access to El Guayacan
and La Cueva. Dr. David Stephan, N. C. State University, kindly identified homopterans.

LITERATURE CITED

Barash, D. P. 1974. An advantage of winter flocking in the Black-capped Chickadee,
Parus atricapillus. Ecology 55:674-676.

Bond, J. 1956. Check-list of birds of the West Indies. Acad. Nat. Sci., Philadelphia.
Cody, M. L. 1971. Finch flocks in the Mohave Desert. Theoret. Pop. Biol. 2:142-158.
Eaton, S. W. 1953. Wood warblers wintering in Cuba. Wilson Bull. 65:169-174.
Gleason, H. A. and M. T. Cook. 1926. Plant ecology of Porto Rico. Scientific Survey
of Porto Rico and the Virgin Islands, 7. N.Y. Acad. Sci.

Horn, H. S. 1966. Measurement of “overlap” in comparative ecological studies. Am.
Nat. 100:419-424.

Janzen, D. H. 1973. Sweep samples of tropical foliage insects: effects of seasons,

vegetation types, elevation, time of day, and insularity. Ecology 54:687-708.
Kepler, C. B. and A. K. Kepler. 1970. Preliminary comparison of bird species diver-
sity in Luquillo and Guanica forests. E-183 to E-191, in A Tropical Rain Forest (H.
T. Odum, ed.), AEG, Oak Ridge, Tennessee.

Krebs, J. R., M. H. MacRoberts and J. M. Cullen. 1972. Flocking and feeding in the
Great Tit Parus major — an experimental study. Ibis 114:507-530.

Lack, D. and P. Lack. 1972. Wintering warblers in Jamaica. Living Bird 11:129-153.
Leck, C. F. 1972. Observations of birds at Cecropia trees in Puerto Rico. Wilson Bull.
84:498-500.

Morse, D. H. 1967a. Competitive relationships between Parula Warblers and other
species during the breeding season. Auk 84:490-502.

. 1967b. Foraging relationships of Brown-headed Nuthatches and Pine Warblers.

Ecology 48:94-103.

. 1968. A quantitative study of foraging of male and female spruce-woods war-
blers. Ecology 49:779-784.

■ . 1970. Ecological aspects of some mixed-species foraging flocks of birds. Ecol.

Monogr. 40:119-168.

. 1971. The foraging of warblers isolated on small islands. Ecology 52:216-228.

214

THE WILSON BULLETIN • Vol. 90, No. 2, June 1978

. 1973. The foraging of small populations of Yellow Warblers and American

Redstarts. Ecology 54:345-355.

Moynihan, M. 1962. The organization and probable evolution of some mixed species
flocks of neotropical birds. Smithson. Misc. Collect. 143:1-140.

Parnell, J. F. 1969. Habitat relations of the Parulidae during spring migration. Auk
86:505-521.

Powell, G. V. N. 1974. Experimental analysis of the social value of flocking by
Starlings {Sturnus vulgaris) in relation to predation and foraging. Anim. Behav.
22:501-505.

Willis, E. 0. 1966a. Interspecific competition and the foraging behavior of Plain-

brown Woodcreepers. Ecology 47:667-672.

. 1966b. The role of migrant birds at swarms of army ants. Living Bird 5:187-

231.

. 1973. Local distribution of mixed flocks in Puerto Rico. Wilson Bull. 85:75-

77.

RESEARCH DIVISION, NORTH CAROLINA DEPT. OF MENTAL HEALTH, BOX 7532,
RALEIGH 27611 ( PRESENT ADDRESS: AIKEN, SC 29801). ACCEPTED 27 SEPT.
1976.

Wilson Bull., 90(2), 1978, pp. 215-220

DDE RESIDUES AND EGGSHELL THINNING IN
LOGGERHEAD SHRIKES

William L. Anderson and Ronald E. Duzan

Relationships among high DDE concentrations, eggshell thinning, and
population declines in birds have received much attention in recent years
(Ratcliffe 1970, Anderson and Hickey 1972, Cooke 1973, Stickel 1975). Most
severely affected are bird- and fish-eating species of Falconiformes and
Pelecaniformes. DDE, a metabolite of DDT, apparently interferes with de-
position of calcium carbonate in the shell gland, with consequences manifested
as thinned eggshells, increased egg breakage, and production of few young.

In this paper we present evidence that the Loggerhead Shrike ( Lanius
ludovicianus) has suffered much the same fate as many of the large predators
and that the causative agent is possibly p,p'-DDE. Shrike populations have
declined in much of the Midwest ( Mayfield 1949, Petersen 1965, Erdman
1970, Graber et al. 1973). Graber et al (1973:7-8) concluded that “there
apparently were two levels of change in the shrike population of northern and
central Illinois — a relatively slow' decline since about 1900, probably related
to the removal of hedges in many areas, and a very rapid decline to near zero
between 1957 and 1965 from causes unknown. It is possible, of course, that
this change is temporary, but as of 1972 we have seen no sign of recovery by
the shrike population.” Erdman ( 1970:150) speculated that the decline of this
passerine in Wisconsin was caused by pesticides.

METHODS

We conducted our study on low-density populations (mean of 7.4 shrikes counted per
day of driving in April) of Loggerhead Shrikes in southern Illinois (15 counties south
of Cumberland County) in 1971 and 1972. Sixty-nine birds were collected during 4 periods
of the year — January, April, July, and September. Twenty clutches (104 eggs) that we
considered complete were collected in April and May. After being weighed, each egg
was opened by cutting tbe shell along the long axis. Samples of fat (subcutaneous and
visceral combined) were excised from tbe birds and, along with the entire contents of
the eggs, were saved for analysis.

The length and breadth of eggs were estimated to the nearest 0.01 mm with a vernier
dial caliper graduated in 0.05-mm intervals. The air-dried shells were weighed to the
nearest 1 mg. Mean thickness of each shell with membrane attached was estimated to
the nearest 1 fi by taking 6 measurements (2 each at the large end, small end. and equa-
tor) with a micrometer graduated in 0.01-mm intervals. The same techniques were used
to determine the length, breadth, and weight of 36 shrike eggs examined in the Chicago
Museum of Natural History. However, measurements for 49 eggs in other arcliival col-
lections were made to the nearest 0.05 mm and 10 mg. Museum eggs were used only if
their blow holes were ^2 mm and if they were taken from “complete” clutches in south-
ern Illinois prior to 1940.

215

216

THE WILSON BULLETIN • VoL 90, No. 2, June 1978

Table

DDE IN Fat of

Logger II f:\ I) Shrikes Collected in Southern

Illinois in 1971-1972

Age and sex

Statistic

Adult females

Adult males

Juveniles^

April and

July

Number of birds

Mean ± SE

21.04 ± 3.51

37.44 ± 9.07

15.04 ± 8.97

Median

15.00

26.39

9.25

Range

3.07-75.00

<0.01-150.00

5.45-33.33

September and

1 January

Number of birds

Mean ± SE

13.73 ± 9.25

10.59 ± 3.84

11.11 ± 11.11

Median

2.38

3.31

<0.01

Range

<0.01-66.6

<0.01-28.57

<0.01-33.33

^ Sexes combined.

Sixteen shrike nests found in 1972 were not disturbed. Each was revisited at 3- to 7-
day intervals to determine the rate of survival of nests and eggs, and the number of j
fledglings produced per successful clutch.

Samples of fat and egg contents were saponified in a KOH-ethanol solution, put
through a florisil column and, if not sufficiently cleaned, subjected to acetonitrile par- ■
titioning. The samples were then analyzed for p,p'-DDE, dieldrin, and heptachlor
epoxide with a Beckman model GC-4 gas chromatograph equipped with an electron cap- i
ture detector. Columns were packed with 1% EPON 1001 resin and 0.5% Viton A J
fluoroelastomer on a solid support of 100-120 mesh Chromosorb W. The column was
operated at 190 °C with ultra-pure helium as the carrier gas flowing at about 45 cc per
min; the detector temperature was 250 °C. The lower limit of detection was considered
to be 0.01 ppm on a wet-weight basis; recovery was 90%.

Statistical tests used in this study are analysis of variance and linear correlation
(Snedecor 1956:160, 268-270).

RESULTS

Pesticide concentrations.- — Detectable concentrations of p.p'-DDE (here-
after called “DDE”) were present in fat of 88% of the 69 Loggerhead Shrikes i
examined. The frequency of occurrence was 93% for 30 adult females, 86% *
for 28 adult males, and 82% for 11 juveniles. Mean concentrations of DDE
were 21.89 ±3.11 ppm (median = 13.88 ppm) lor all birds, 19.33 ± 3.44
ppm for adult females, 27.85 ± 6.41 ppm for adult males, and 13.96 ± 3.90 ;
ppm for juveniles.

Shrikes collected in April and July — i.e., local breeders and their young —
contained greater concentrations of DDE than did shrikes collected in Sep- I

Anderson and Diizan • DDE AND SHRIKES

217

Table 2

Characteristics of Eggs of Loggerhead Shrikes Collected in Southern Illinois:

1875-1895 AND 1971-1972

Characteristic

Mean ±

1875-1895

1971-1972

F Value

(1)

Length fmm)

24.72 ±0.11(84)"

24.95 ±0.11(61)

2.14

(2)

Breadth (mm)

18.67 ± 0.04(85)

18.52 ±0.07(61)

3.53

(3)

Weight (mg)

251 ±2(83)

244 ± 2 ( 98 )

2.79

(4)

Size Index (1) X i2)

451 ±2(84)

462 ± 3(61)

0.03

(5)

Thickness Index
(3)/(4)

0.544 ± 0.003(83)

0.530 ± 0.005(57)

6.6P

(6)

Thickness (fx)

—

92 ± 0.4(95)

(7)

Weight of whole egg
(g)

—

4.25 ± 0.05(104)

^8)

DDE (ppm)®

—

3.09 ± 0.09 (104)

^ Number of eggs.

2 Significant ( P < 0.05 ) .

2 Median = 1.79, range = 0.48— 34.14.

tember and January (Table Ij. The difference between the means for adult
males was significant (P < 0.05).

The contents of the 104 eggs analyzed contained a mean concentration of
3.09 ppm of DDE (Table 2). A clutch of 6 eggs collected in 1971 had a mean
of 17 ppm, with 1 egg containing a high of 34 ppm. Dieldrin and heptachlor
epoxide were not detected in the eggs or the samples of fat.

Physical characteristics of eggs. — The mean value for the shell thickness
index was 2.57% less for shrike eggs collected in 1971 and 1972 than for eggs
of this species collected between 1875 and 1895 (Table 2). This difference
was significant (P < 0.05). Mean values for other physical characteristics —
length, breadth, weight, and size index — did not differ significantly between
the recently collected and older eggs.

Linear correlation indicated that a negative relationship existed between
concentrations of DDE and the thickness of shells for the recent eggs: Y =
92.610-2.412 logio X, r = -0.208 with 93 df (P<0.05). Correlations be-
tween concentrations of DDE and other physical characteristics were not
significant.

Rothstein (1972) found that the number of eggs in the clutch and the
degree of embryonic development influenced eggshell thickness in Cedar
Waxwings (Bomhycilla cedrorum) . This was not true of the shrike eggs col-
lected in 1971 and 1972. In clutches with ^5 eggs, mean ± SE eggshell thick-
ness was 92.4 ± 0.8 [x for 17 eggs without development and 94.0 ± 1.5 fx for 4
eggs with development. In clutches of 6-7 eggs, mean thickness was 89.9 ± 6

218

THE WILSON BULLETIN • Vol. 90, No. 2, June 1978

jx for 12 eggs without development and 92.0 ± 0.5 /x for 57 eggs with develop-
ment. None of the differences among these means was significant ( F = 1.86
with 3 and 86 df ) .

Nest success. — As determined by Mayfield’s (1961) day exposure method,
survival of nests studied in 1972 was 79% during incubation (n = 13), 91%
during the nestling period (n = 13j, and 72% from start of incubation to
fledging. Survival of eggs was 75% during incubation (n = 74) and 83%
during the period of hatching ( n = 54) . Survival of young during the nestling
period was 88% (n = 57). Thus, 55% of the eggs present at the beginning
of incubation produced young that eventually fledged. A mean of 3.9 young
fledged per successful nest ( n = 9) .

DISCUSSION

Data obtained during this study strongly suggest that Loggerhead Shrikes
in Illinois have acquired appreciable amounts of DDE and that eggshell
thickness has been adversely affected. However, the shrikes were not as
severely contaminated with DDE as some raptorial and piscivorous species —
birds well known for the eggshell thinning syndrome, poor reproductive suc-
cess, and population declines. For example, mean concentrations (wet-w eight
basis) in Peregrines {Falco peregrinus) in Alaska were 38.2 ppm in fat of
juveniles, 622.0 ppm in fat of adults, and 12.48 ppm in eggs (Cade et al. 1968:
175). Mean concentrations in eggs from North America and western Europe
were 8.6 ppm for 16 species of Falconiformes and 17.7 ppm for 5 species of
Pelecaniformes, as calculated from data presented by Stickel (1973:260-267).

The shrike’s high position in the food pyramid is almost certainly the over-
riding factor leading to DDE accumulation in the species. The recent finding
by Graber et al. (1973:12), who examined stomachs of the birds we analyzed
for pesticides, that shrikes frequently consume ground beetles (Carabidae)
is particularly relevant. Because of the predaceous habits of these insects,
they themselves might be expected to accumulate pesticides, which would be
passed on to shrikes and other predators that feed on the beetles.

The relationship between eggshell thinning and high DDE concentrations in
shrikes parallels the well-documented eggshell thinning syndrome in the Pere-
grine, Bald Eagle i Haliaeetus leucoccphalus ) , Brown Pelican [ Pelecanus oc-
cidentalis), and certain other birds (Ratcliffe 1970, Anderson and Hickey
1972, Faber and Hickey 1973, Blus 1974). Like these species, the eggshell
thinning in shrikes occurred concurrently with declining or reduced popula-
tions. However, our data on nesting success reveal that shrikes were highly
successful in producing fledglings in the low-density population in southern
Illinois in 1972. Graber et al. (1973:9) reported similar findings for shrikes
in central Illinois in 1958-64 (population now extirpated) and in south-

Anderson and Duzan • DDE AND SHRIKES

219

eastern Illinois in 1967. Broken eggs, crushed embryos, or other indications
of atypical egg mortality were not detected I Richard R. Graber, pers. comm.).

We conclude that the factor or factors that caused the decline of the Log-
gerhead Shrike population in Illinois were more closely associated with sur-
vival of fledged juveniles or adults than with reproduction. We have not
demonstrated that the causative factor was DDE. Nevertheless, suspicion
can be directed toward this environmental toxicant because ( 1 1 it has con-
taminated the shrike population and (2 I a relationship exists between it and
the malfunction of at least 1 physiological process — eggshell thickness — in
the species.

SUMMARY

Investigations in southern Illinois in 1971 and 1972 suggest that the Loggerhead
Shrike has been contaminated with DDE and that the species has experienced eggshell
thinning. Mean concentrations of DDE were 21.89 ppm in fat of 69 birds and 3.09 ppm
in the contents of 104 eggs. A negative correlation was found between concentrations of
DDE and eggshell thickness, and the mean value for the shell thickness index was 2.57%
less for eggs collected during the study than for eggs in archival collections. However,
nesting success was high, suggesting that the factor — DDE or other — causing the recent
decline of the shrike population in Illinois was more closely associated with survival of
I fledged juveniles or adults than with reproduction,
i

ACKNOWLEDGMENTS

Appreciation is extended to the following personnel of the Illinois Natural History Sur-
vey: to J. W. Seets for assistance in collecting shrikes, locating nests, and dissecting
1 specimens; to G. C. Sanderson and H. C. Schultz for editorial help; and to R. R. Graher
' for critically reading the manuscript. E. E. Klaas. Patuxent Wildlife Research Center,

1 Laurel, Maryland, kindly provided measurement data on eggs in archival collections.

LITERATURE CITED

' Anderson, D. W. and J. J. Hickey. 1972. Eggshell changes in certain North American
birds. Proc. XV Int. Ornithol. Congr. : 514-540.

Blus, L. j., a. a. Belisle, and R. M. Prouty. 1974. Relations of the Brown Pelican
i to certain environmental pollutants. Pestic. Monit. .1. 7:181-194.

i Cade, T. J., C. M. White, and J. R. Haugii. 1968. Peregrines and pesticides in Alaska,
i Condor 70:170-178.

; Cooke, A. S. 1973. Shell thinning in avian eggs by environmental pollutants. Environ,
i Pollut. 4:85-152.

Erdman, T. C. 1970. Current migrant shrike status in Wisconsin. Passenger Pigeon
32:144-150.

Faber, R. A. and J. J. Hickey. 1973. Eggshell thinning, chlorinated hydrocarbons, and
mercury in inland aquatic bird eggs, 1969 and 1970. Pestic. Monit. J. 7:27-36.

^ Graber, R. R., J. W. Graber, and E. L. Kirk. 1973. Illinois birds: Laniidae. Illinois
Nat. Hist. Surv. Biol. Notes 83.

220

THE WILSON BULLETIN • VoL 90, No. 2, Jane 1978

Mayfield, H. 1949. Nesting season. Middlewestern prairie region. Audubon Field
Notes 3:237-238, 240.

— . 1961. Nesting suecess ealculated from exposure. Wilson Bull. 73:255-261.

Petersen, P. C., Jr. 1965. Spring migration. Middlewestern prairie region. Audubon
Field Notes 19:480-482.

Katcliffe, D. a. 1970. Changes attributable to pesticides in egg breakage frequency
and eggshell thickness in some British birds. J. Appl. Ecol. 7:67-115.

Kotiistein, S. 1. 1972. Eggshell thickness and its variation in the Cedar Waxwing.

Wilson Bull. 84:469-474.

Snedecor, G. W. 1956. Statistical methods. Iowa State Univ. Press, Ames.

Stickel, L. F. 1973. Pesticide residues in birds and mammals. Pp. 254-312 in En-
vironmental Pollution by Pesticides (C. A. Edwards, ed.). Plenum Press, New York.

Stickel, W. H. 1975. Some effects of pollutants in terrestrial ecosystems. Pp. 25-74 in
Ecological Toxicology Research (A. D. McIntyre and C. F. Mills, eds.). Plenum Pub-
lishing Corp., New York.

SECTION OF WILDLIFE RESEARCH, ILLINOIS NATURAL HISTORY SURVEY, URBANA

61801 (present address WLA: division of wildlife resources, ILLI-
NOIS DEPT. OF CONSERVATION, SPRINGFIELD 62701.) ACCEPTED 10 JAN.
1977.

Wilson Bull, 90(2), 1978, pp. 221-238

TREE SPECIES USED BY BIRDS IN LOGGED AND
UNLOGGED MIXED-CONIEEROUS EORESTS

Kathleen E. Franzreb

A number of studies have examined the possible ways birds use their hal)-
itat to permit optimum exploitation and to allow coexistence of different
species. Habitat partitioning may occur if, for example, each species pos-
sesses a characteristic foraging height distribution ( Hartley 1953, Morse
1968, Diamond 1973). Birds may also divide the vegetation into various
horizontal levels (Colquhoun and Morley 1943, MacArthur 1964, Baida
1969). Individuals of a given species may also primarily restrict their ac-
tivities to specified zones in the tree such as the trunk or outer branches
( Hartley 1953, MacArthur 1958, Morse 1967a, 1967b, Stallcup 1968, Stur-
man 1968, Koch et al. 1970 j . There is evidence that some avian species do
have preferences for certain plant species (Hartley 1953, Morse 1967b,
Baida 1969, Willson 1970, Jackson 1970, Austin and Smith 1972). These
: factors by acting individually or in concert may interact to produce habitat
partitioning for the various avian species in the community.

I It has been suggested that plant species diversity has little effect on a
I community’s bird species diversity which instead, is dependent upon foliage
1 height diversity, at least in homogeneous environments (MacArthur and
' MacArthur 1961). It is the vegetation profile which determines bird species
I diversity and not the number of plant species present. Yet, the nature of the
I vegetation may be important in a heterogeneous environment. MacArthur
I (1964) found that in the complex environment of the Chiricahua Mountains,
Arizona, birds “apparently used more than just profile in selecting suitable
habitats; presumably the addition of nest holes and water, the change from
j oak to pine and from sparse to dense foliage, all made significant changes
I in the acceptability of the habitat for many species.” It appears that birds do
not regard all trees of the same height and profile, belonging to different
j species, as being equally desirable for activities such as foraging and nesting.

' The purpose of this investigation was to determine if birds select certain
' tree species in a mixed-coniferous forest, and if so, to what extent the use
I of a logged area differs from that of a virgin forest. Results from this study
I may suggest some guidelines for future timber harvesting practices in the
I Southwest.

STUDY AREA AND METHODS

^ Study area. — The study site was located on tlie Willow Creek watershed (202 ha logged;
131 ha unlogged), approximately 80 km south of Springerville on the Apache-Sitgreaves

221

222

THE WILSON BULLETIN • Vol. 90, No. 2, June 1978

National Forest, Greenlee Co., White Mountains, Arizona. It is a U.S. Forest Service
experimental watershed ranging in elevation from 2682 to 2804 m.

Vegetation. — Sampling of trees with dhh (diameter at breast height) ^7.6 cm in the
unlogged and lumbered areas was conducted using the plotless point-quarter method
(Cottam and Curtis 1956). One-hundred stations (400 trees) were sampled in a 15.5
ha study plot in each area. For the tree in each quadrat closest to the center stake, the
following data were taken: tree species; tree height; dhh; and distance from the center
of the trunk to the stake. Quaking aspen i Populus tremuloides) and snags (dead trees)
had not been removed when part of the watershed was logged in the summer of 1972.

Foliage volume. — Use of the available live foliage volume was analyzed by estimating
the amount of foliage for each tree species in the modified and unaltered habitats. Data
for these analyses were collected simultaneously with those of the point-quarter measure-
ments. Tree species, tree height, height to the first live branch, length of the longest
branch, and distance from the center of the trunk to the first live foliage on the longest
branch were recorded for each tree sampled.

An estimate of live foliage present for a given tree species was determined by cal-
culating the total amount of foliage for each tree and then subtracting from this the
value of the dead volume ( the inner area of the tree which was devoid of live needles
or leaves). To obtain volume in terms of mVha for a particular tree species, I divided
the live volume for each species by the number of trees of the given species sampled.
Next, I multiplied this by the absolute density of the species sampled. The result is live
foliage volume (mVha) of the given tree species.

The actual formulae used to estimate volumes for a given tree were:

Live foliage volume for spruces iPicea spp.), firs (Abies spp.), and Douglas-fir
(Pseudotsuga menziesii) tt/3 (ro"Ao - ri^/ii) where hi — h) - {ro - n) i

Live foliage volume for pines (Pinus spp.) = TrirJ'ho- ri'hi) where hi = ho- (ro-ri) j
Live foliage volume for quaking aspen = 4/3 7r(ro^- ri )

In all cases, ro represents the length of the longest branch and ri is the distance from the '
center of the trunk to the beginning of the live vegetation on the longest branch, ho rep-
resents the height of the tree’s live vegetation (i.e., total tree height - height to first ;
branch), hi is the height of the portion of the tree containing branches which have dead ij
foliage, minus the height to the first branch.

Avian tree species selection. — When a bird was observed in a tree, the following data
were recorded: bird species, tree species, and activity (singing, foraging, observing, or
resting). Data were collected at all daylight hours throughout the summers of 1973 and
1974 by systematically traversing the study plots along established parallel transect lines.
One observation per bird was recorded and the number of such observations is indicated
by N. I obtained 4868 total observations in the unharvested site and 4964 in the modified
habitat.

Preferences for certain tree species were noted by comparing frequency of use of a ,
particular tree species with its percent availability in the habitat as calculated from
foliage volume data. Relative density of snags as determined from point-quarter data
was used for comparative purposes as no foliage was found on them.

Data were analyzed separately for the Yellow-bellied Sapsucker iSphyrapicus varius), ]
Mountain Chickadee ^ Parus gambeli), Ruby-crowned Kinglet (Regulus calendula), I
Yellow-rurnped Warbler ( Dendroica coronata) , and Gray-headed Junco {Junco cani- j
ceps). Tliese 5 species were selected because they are representative of hole, open-cup,
and ground nesting species and because they were present in sufficient densities in both i

Franzreb • TREE SPECIES USED BY BIRDS

223

habitats to afford adequate sample sizes. Data for the entire avifauna included these
as well as all other species. Niche breadth was calculated for each species using the fol-
lowing formula (Shannon 1948) : B = -^pi In pt where B is the habitat niche breadth
and Pi is the proportion of observations occurring in the tree species.

Avian densities. — Species densities were determined using the spot-map method (Wil-
liams 1936) whereby a 15.5 ha grid pattern was established in the unlogged and har-
vested areas using plastic flagging placed at 25 m internals along 9 parallel lines, each
390 m in length, and 50 m apart. 1 labelled each flag with a number corresponding to the
transect line and a letter corresponding to the distance traveled from the beginning of
the line. Censusing was conducted 6 times per month (June, July, and August) in each
habitat.

RESULTS

Vegetation. — In the unharvested area total tree density was 626.2 trees per ha
versus 167.7 in the harvested site (Table 1 j. In both habitats, Douglas-fir had
the highest density as well as the highest importance value. In the logged
area snags and quaking aspen, neither of which were removed during logging,
were of considerable importance. The category “snags” contained representa-
tives of every tree species.

Foliage volume. — Ponderosa pine [Pinus pondersoa) (35.9%), followed
closely by southwestern white pine [Pinus strobiformis) (35.3%), and
Douglas-fir (17.6%) had the greatest foliage volume in the unharvested site
(Table 1 ) . However, in the logged habitat quaking aspen contained the ma-
jority of total foliage volume available (53.1%). Southwestern white pine
(12.6%) and Douglas-fir (11.0%) comprised considerably less foliage volume
than did aspen.

Avian tree species selection. — I used iSeyman’s (1949) statistical test to
evaluate differences between actual avian tree species use and the expected
number of observations based upon the foliage volume availability of the var-
ious tree species. The following results were all statistically significant at the
P ^ 0.05 level unless otherwise indicated.

Tree species preferences for the entire avifauna in the unaltered site (Fig.
1) indicated significant preferences for Douglas-fir, white fir [Abies con-
color), and Engelmann spruce {Picea engelmanni) . However, certain tree
types were not used to the extent to which they were available such as pon-
derosa pine, southwestern white pine, and snags. There was no significant
difference between the volume of aspen available and bird use either by the
total avifauna or by the 5 individual avian speices in the unlogged habitat.

In the lumbered area, avian use of aspen was significantly less than ex-
pected based on the volume of foliage present for both the total avifauna and
the 5 selected bird species. Douglas-fir and Engelmann spruce were strongly

224

THE WILSON BULLETIN • Vol. 90, No. 2, June 1978

Table

Tree Species Density, Importance Value, and Foliage Volume

Unlogged area

Tree density
(#/ha)

Importance

value*

Foliage volume

Tree species

(m^/ha)

Percent

Ponderosa pine

112.7

67.8

40910.4

35.9

Southwestern white pine

109.6

46.7

40253.3

35.3

Alpine fir

3.1

1.5

181.7

0.2

Douglas-fir

194.1

92.3

20000.5

17.6

White fir

51.7

24.5

4305.5

3.8

Blue spruce

12.5

5.2

552.6

0.5

Englemann spruce

31.3

13.0

2213.3

2.0

Quaking aspen

50.1

20.3

5565.9

5.0

Snag (dead tree)

61.1

28.7

—

Total

626.2

300.0

113984.1

100.0

Logged area

Tree density
(#/ha)

Importance

value*

Foliage volume

Tree species

(m^/ha)

Percent

Ponderosa pine

4.6

16.3

1069.2

7.0

Southwestern white pine

8.8

14.5

1921.8

12.6

Alpine fir

13.0

20.8

497.0

3.3

Douglas-fir

42.3

64.0

1679.2

11.0

White fir

19.7

30.5

544.9

3.6

Blue spruce

9.6

13.9

421.0

2.8

Engelmann spruce

19.3

31.0

1030.1

6.8

Quaking aspen

29.3

51.2

8105.6

53.1

Snag (dead tree)

21.0

57.8

—

Total

167.7

300.0

15269.9

100.2

* Importance value is the sum of relative density, relative dominance, and relative frequency.

preferred, whereas ponderosa pine, southwestern white pine, and aspen were
seldom used.

In Both habitats the 5 avian species preferred spruces and firs and used the
pines, aspen, and snags less than expected ( Figs. 2-6 j. There were several
exceptions to these generalizations. For example. Yellow-bellied Sapsuckers
(Fig. 2j frequently foraged on snags in both habitats. No significant differ-
ence between use and availability was found for Douglas-fir or white fir in
the unlogged site or for blue spruce and Engelmann spruce in the logged site.

It may he argued that since Yellow-bellied Sapsuckers confine most of their
activities to tree trunks, a comparison of sapsucker use to tree species density

Franzreb • TREE SPECIES USED BY BIRDS

225

TREE SPECIES PREFERENCES
RLE BIRDS
UNLOGGED RREH

TREE SPECIES PREFERENCES
RLL BIRDS
LOGGED RRER

Fig. 1. Tree species preferences of all birds in the iinlogged and logged areas. Plain
bar is % of avian use. Crosshatched bar is % of foliage volume for the given tree species.
Hatched bar is relative density of snags. Tree species were abbreviated as follows: PP —
ponderosa pine; SWWP — southwestern white pine; AF — alpine fir; DF--Douglas-fir;
WF — white fir; BS — blue spruce; ES — Englemann spruce; and ASP — (juaking aspen.

PERCENT PERCENT

0.00 15.00 30.00 45.00 60.00 0^.00 15.00 30.00 45.00 60.00

226

THE WILSON BULLETIN • Vol. 90, No. 2, June 1978

TREE SPECIES PREFERENCES
UNLOGGED RREfl
YELLOW-BELLIED SflPSUCKER

□ RVIRN USE
@ FOLIRGE VOLUME

^ RELRTIVE DENSITY

TREE SPECIES

N = 142

TREE SPECIES PREFERENCES
YELLOW-BELLIED SflPSUCKER
LOGGED AREA

PP SWWP RF OF WF BS ES flSP SNRG

TREE SPECIES

N= 6 9 3

□

RVIRN USE
FOLIRGE VOLUME
RELRTIVE DENSITY

Fig. 2. Tree species preferences of the Yellow-l)eIIied Sapsucker in the unlogged and
logged areas.

PERCENT PERCENT

0..00 15.00 30.00 45.00 60.00 0,^2^ 15.00 30.00 45.00 60.00

Franzreb • TREE SPECIES USED BY BIRDS

227

TREE SPECIES PREFERENCES
MOUNTAIN CHICKADEE
UNLOGGED AREA

□ RVIflN USE
^ FOLIRGE VOLUME
^ RELRTIVE DENSITY

PP SWWP RF OF WF BS ES RSP SNRG

TREE SPECIES

N= 9 91

TREE SPECIES PREFERENCES
MOUNTAIN CHICKADEE
LOGGED AREA

TREE SPECIES

N= 59 9

□

RVIRN USE
FOLIRGE VOLUME
RELATIVE DENSITY

Fig. 3. Tree species preferences of tlie Mountain Chickadee in the unlogged and
logged areas.

PERCENT PERCENT

0.00 15.00 30.00 45.00 60.00 Oi'OO 15.00 30.00 45. 00 60.00

228

THK WILSON miLLETIN • Vul. 90, .Yo. 2, June 1978

TREE SPECIES PREFERENCES
RUBY-CROWNED KINGLET
UNLOGGED AREA

□ flVIflN USE
^ FOLIRGE VOLUME
^ RELATIVE DENSITY

PP SWWP RE DF WF BS ES RSP SNRG

TREE SPECIES

N= 7 7 9

TREE SPECIES PREFERENCES
RUBY-CROWNED KINGLET
LOGGED AREA

PP SWWP RF OF WF BS ES RSP SNRG

TREE SPECIES □ use

N= 7 5 7 ^ FOLIRGE VOLUME

^ RELATIVE DENSITY

Fig. 4. Tree species preferences of the Ruby-crowned Kinglet in the unlogged and
logged areas.

Franzreb • TREE SPECIES USED BY BIRDS

229

TREE SPECIES PREFERENCES
YELLOW-RUMPED WfiRBLER
UNLOGGED RRER

TREE SPECIES

N= 780

TREE SPECIES PREFERENCES
YELLON-RUMPED WRRBLER
LOGGED RRER

Fig. 5. Tree species preferences of the Yellow-runiped Warbler in the unlogged and
logged areas.

230

THE WILSON BULLETIN • VoL. 90, No. 2, June 1978

TREE SPECIES PREFERENCES
GRflY-HEfiDED JUNCG
UNLOGGED RRER

TREE SPECIES

N= 280

TREE SPECIES PREFERENCES
GRflY-HEfiDED JUNCO
LOGGED AREA

Fig. 6. Tree species preferences of the Gray-headed Junco in the unlogged and logged

areas.

Franzreb • TREE SPECIES USED BY BIRDS

231

Table 2

Niche Breadth Based on Tree Species Use

Bird species

Unlogged

Logged

Yellow-bellied Sapsucker

2.00

1.93

Mountain Chickadee

1.74

1.90

Ruby-crowned Kinglet

1.71

1.4f)

Yellow-rumped Warbler

1.91

1.96

Gray-headed Junco

1.72

1.90

* Niche breadth = In pi

might be more appropriate than a comparison to foliage volume data since
presumably this species would he less dependent on the volume and type of
foliage present than would he, for example, a foliage-gleaning species. How-
ever, the density of individual tree species is positively correlated with fo-
liage volume per tree species (r = +.69 unlogged area, r — +.41 logged site).
Results for the sapsucker were graphed in a consistent manner w ith the repre-
sentation of the other 4 species and with the total avifauna.

For the Ruby-crowned Kinglet there was no significant difference between
frequency of use and proportion of foliage volume present in alpine fir and
blue spruce in the lumbered section (Fig. 4). The kinglet was rarely observed
on snags.

The Yellow-rumped Warbler in the unharvested area (Fig. 5) visited pon-
derosa pine and southwestern white pine more frecjuently than did the other 4
avian species examined in detail. In the modified site, aspen was also fre-
(luently used, but far less than expected on the basis of tree availability.

Gray-headed Juncos in the unaltered habitat used alpine fir and aspen in pro-
portion to their availability ( Fig. 6). Tree use in the harvested area could not
be associated w ith tree availability.

The Ruby-crowned Kinglet had a higher niche breadth value (Table 2) in
the unlogged area than in the modified site. In contrast, the Mountain Chick-
adee and Gray-headed Junco showed higher values in the logged area. Niche
breadths for both the Yellow-bellied Sapsucker and Yellow-rumped Warbler
were similar in the modified and virgin forests. For the unaltered habitat,
the Yellow-bellied Sapsucker iB = 2.001 had the highest niche breadth value
followed by the Yellow-rumped Warbler (B — 1.91). These 2 species also had
the highest values for the modified area, B = 1.93 and B = 1.96, respectively.

Although the Mountain Chickadee and Ruby-crowned Kinglet displayed
strong preferences for the same tree species there were differences in terms of
proportion of use. In the unmodified site the Mountain Chickadee used

232

THE WILSON BULLETIN • VuL 90, No. 2, June 1978

Bird Densities In Logged and

Table 3

Unlogged Mixed-Coniferous Forests (#/40 iia)

Bird species

1974

1973

Logged

Unlogged

Logged

Unlogged

Yellow-bellied Sapsucker

15.8

10.6

20.5

10.2

Mountain Chickadee

11.8

44.7

30.8

58.9

Ruby-crowned Kinglet

42.1

71.0

23.1

74.4

Yellow-rumped Warbler

100.0

131.6

76.9

89.8

Gray-headed Junco

76.3

31.6

74.4

51.3

Total avifauna

544.0

632.9

758.0

865.9

Douglas-fir 40.9% and Engelmann spruce 13.6% of the time. However, the
Ruby-crowned Kinglet frequented Engelmann spruce 45.8% and Douglas-fir
only 22.7% of the time. A slightly different situation occurred in the logged
site where the Mountain Chickadee was less specialized in its tree species use,
selecting Douglas-fir and Engelmann spruce a total of 48.1%, whereas the
Ruby-crowned Kinglet relied heavily on these 2 tree species (82.6% of its to-
tal foraging observations).

Avian densities. — The unmodified habitat supported a considerably larger
avian community (Table 3) than did the logged site during the breeding
season of each year. The Mountain Chickadee, Ruby-crowned Kinglet, and
Yellow-rumped Warbler were more numerous in the virgin forest than in the
lumbered site. Gray-headed Juncos were much commoner and Yellow-bellied
Sapsuckers were slightly more numerous in the harvested site than in the un-
logged area.

DISCUSSION

There are various possible explanations for avian selection of certain tree
species while not using others. Factors affecting tree selection include food
abundance, availability, and quality. Foliage may be important for birds in
that it protects them from predators and inclement weather conditions and
shelters the nest sites of numerous species.

In examining foraging behavior of English titmice {Pams spp.) Hartley
(1953) found that although every tree species was used, some were far more
intensively used. Tree species selection, in addition to height distribution and
the tendency to search for food in different parts of the trees, permitted 5
species of titmice to occupy the same habitat. During periods of superabun-
dant food supply, the ecological distinctiveness in feeding behavior disap-
peared among the 5 congeners, indicating the importance of competition for

Franzreb • TREE SPECIES USED BY BIRDS

233

food in determining foraging behavior ( Hartley 1953 j . Studies by Palm-
gren (1930), Kluijver (1951), Gibb (1954), and Jackson (1970) also noted
avian selection of certain tree species. However, Brewer (1963) stated that
the Black-capped \ Pants atricapillus) and Carolina chickadees (P. carolinen-
sis) appeared to use the vegetation in relation to its abundance in the habitat
in Illinois. Perhaps no habitat partitioning was evident in these species
because they are primarily allopatric; moreover. Brewer suggests that competi-
tive interactions may be a factor in producing this allopatry. Habitat parti-
tioning achieved, in part, by tree species selection in Willow Creek was prob-
ably of more importance to some species than to others. No 2 species had
exactly the same tree species preferences. Some potential competitors such as
the Mountain Chickadee and Ruby-crowned Kinglet spent different propor-
tions of their time in the same tree species. In the absence of data on abun-
dance and locations of the food supply, it is not possible to state with cer-
tainty that competition for food was responsible for this habitat partitioning,
j In addition to tree species selection a variety of other possible mechanisms
i such as within-tree location preferences (Franzreb 1976), tree height selec-
tion, and foraging behavior (Franzreb 1975) were involved in habitat
segregation.

! During the nesting season the majority of birds in Willow Creek are in-
sectivorous. The availability and type of food source may influence which
trees birds prefer. Southwood (1961) found that the number of insect species
associated with given tree species varied in Britain. Among genera of trees
I in his study (and which occurred in Willow Creek), poplars (Populus), pines
I iPinus), spruces iPicea), and fir {Abies) had 97, 91, 137, and 15 insect
j species respectively. Overall insect abundances were not derived, hence, it
j was not known which tree species harbored the greatest densities of insects,
i Birds may select trees with only a few insect species if such prey items occur
I in sufficient numbers to make exploitation efficient. Insect abundance and
1 number of species may also have varied among the tree species in Willow
; Creek.

I The amount of vegetation present may influence insect numl>ers as evi-
denced by the number of foraging observations in heavy foliage. Successful
; foraging, however, is not solely dependent on the number of prey items per
I tree, but is more closely related to the density of insects (number per unit
I foliage volume). Search time required between successful captures, flying
time between foraging sites, handling time per prey item, and the individual’s
degree of prey specificity may determine a bird’s competency in exploiting a
j given resource. Those individuals selecting the portions of the habitat in
, which it is possible to achieve the highest degree of foraging success will have
an advantage.

234

THE WILSON BULLETIN • Vol. 90, No. 2, June 1978

In the unlogged study area in Willow Creek, ponderosa pine and south-
western white pine, the tree species which contained the greatest amount of
foliage volume were used far less than other tree species containing smaller
amounts of foliage such as Douglas-fir and Engelmann spruce. Also, for the
timber harvested area quaking aspen comprised 53% of the total available
foliage and accounted for less than 15% of the total avian observations. How-
ever, foliage volume calculations were based on formulae which did not
consider that pine and aspen leaves or needles are much less dense than those
of spruces and firs. Therefore, the foliage per unit volume probably was
much higher for the latter tree species than for the former. Hence, the dis-
parity between avian use and calculated available foliage for some tree species
may not be as great as shown.

Possibly the majority of birds infrequently visited pines and aspen be-
cause it was not energetically economical to use them. Perhaps search time
was so prolonged that in all but the choicest sections of these trees, exploita-
tion was impractical. Increased exposure to predators and inclement weather
resulting from the openness of the vegetation may also have discouraged birds
from using these species.

Snags harbor a variety of insects, many lying between the crevices on the
bark or under the bark’s surface, thus necessitating special morphological
adaptations such as in the tongue or in the length and shape of the bill, to pro-
vide accessibility to prey. The many typically foliage-gleaning species in Wil-
low Creek lack the necessary adaptations to extensively use snags but are
adapted for exploiting live trees. Snags also harbor a different insect fauna
than found in live trees (Anderson 1960). Difficulty in securing food as well
as the type of insect prey available in snags, may have prevented or dis-
couraged many birds, particularly foliage gleaners, from extensively using
them.

Leaf morphology and size may influence the degree of avian use of a
given tree species. The large leaf size of quaking aspen make it difficult for
the majority of these birds, particularly the smaller passerines, to perch on
an aspen branch or twig and reach the middle and outer portions of the
leaves which may harbor insects. Hovering forms such as the Ruby-crowned
Kinglet may encounter difficulty since aspen leaf movement occurs with
even slight breezes. Further, aspens may not support similar insect densities
and species found in coniferous trees.

Niche breadth values derived from tree species use data for individual
bird species indicated differences existed between the unlogged and modified
sites. In this context, I used the spatial model for the niche proposed by
Hutchinson (1958) and expanded by Slobodkin (1962), Levins (1968), and
MacArthur (1968). Avian niche breadth was considered as the degree of

Franzreb • TREE SPECIES USED BY BIRDS

235

diversity in tree species selection. It is the inverse of the degree of ecological
specialization of a particular species within a given habitat (Levins 1968).
Niche breadth does not indicate anything about tree species availability and
is only suggestive of the evenness of tree species use for a particular bird
species.

It is possible that very stereotyped species in the logged site which did not
occur in sufficient numbers to be studied, were among the most affected by
habitat modification. For example, logging in effect eliminated the Brown
Creeper [Certhia familiaris ) , thus making it impossible to compare its forag-
ing behavior in the 2 study areas. Perhaps those species present in sufficient
numbers in both habitats to afford an adequate sample size, were the most
plastic species, yet 1 of the 5 appeared to be fairly stereotyped in its foraging
behavior (Ruby-crowned Kinglet). Stereotyped species should show a re-
duction in density and/or niche breadth in the modified site since they, pre-
sumably, restricted their activities to only the most useful tree species. This
was true for the Ruby-crowned Kinglet. Such species may have enlarged
territories in order to be assured of access to the minimum essential number
of preferred trees. Thus, the number which the logged habitat could support
would be reduced.

A more generalized species which maintained its density in the logged site
map show a concomitant increase in niche breadth since it was possible to
switch to normally infrequently visited tree species. A species might undergo
a reduction in density in the modified site if either its niche breadth value
was approximately the same in the 2 study sites (i.e., it could not become
more generalized) or if the niche breadth value increased. The former situa-
tion was found for the Yellow-rumped Warbler and the latter for the Moun-
tain Chickadee. Apparently the logged habitat was incapable of supporting
as many Mountain Chickadees as the unmodified site. Perhaps this resulted
from an insufficient amount of foliage volume of the favored tree species to
fulfill all their requirements.

If a species, such as the Gray-headed Junco, achieves a higher density in
the altered site than in the virgin forest, it may he fairly plastic and hence,
i exhibit an increase in niche breadth. Gray-headed Juncos were more numer-
I ous and had a higher niche breadth value in the logged than unlogged site.

, This junco, primarily a ground foraging species, is not as specific in its tree
species selection as are foliage-gleaners. The Juncos also used slash (logging
debris) which was abundant following lumbering. Slash provided an addi-
tional substrate which probably contributed to their higher density there.

■ Some species may have concentrated on the preferred tree species in order
to acquire essential resources such as food, or perhaps it was a useful

236

THE WILSON BULLETIN • VoL 90, No. 2, June 1978

strategy to further segregate the habitat among species in the logged site
which may have been a more limited environment. Others may have become
more generalized in order to take advantage of a larger portion of the habitat.
Apparently avian species adjusted in various ways to the reduction in foliage
volume and the reapportionment of the available foliage to different tree i
species in the harvested environment. |

Future management decisions pertaining to timber harvesting should con- ,
sider the heavy use of spruces and firs by birds. A decision to remove a sub-
stantial proportion of pines and aspen even though they are not frequently j
used by the avifauna, would adversely affect a number of species which rely l
upon these trees for nesting such as the Yellow-bellied Sapsucker and ;
Warbling Vireo {Vireo gilvus) which nest in aspen, and for foraging such |
as the Grace’s Warbler (Dendroica graciae) which forages in pines. I recom- i
mend that areas not be logged as heavily as Willow Creek which underwent
a moderately-heavy overstory removal (removal of most of the trees forming I
the forest canopy). The majority of snags should not be harvested as they i
serve several significant functions such as providing nest sites for numerous •
cavity-nesting species. '

SUMMARY

Variation in avian selection of tree species in a community which had undergone an
overstoiy' removal form of timber harvesting was compared to a virgin, mixed-coniferous i
forest, in the White Mountains, Arizona, during the summers of 1973 and 1974. Tree j
species preferences for all birds observed indicated Douglas-fir, white fir, and Engelmann fI
spruce were the most frequently visited species in both habitats and were used in both the 'I
unmodified and logged areas in excess of the proportion of foliage volume they contained ,
in the entire habitat. Ponderosa pine and southwestern white pine were frequented less-|
than expected on the basis of availability. Although aspen constituted over 50% of the
available foliage in the harvested habitat, birds did not appear to compensate for the ;
reduction in density of firs and spruces by increasing their use of it. Whereas some J
species in the modified environment, such as the Mountain Chickadee, became more j
generalized and therefore less selective as to tree species, the Ruby-crowned Kinglet ap- ;
parently l)ecame more restricted, and hence, more specialized in tree species preferences, i
Use of quaking aspen, the only species not removed during harvesting, and snags (dead ;
trees) was higher in the modified than in the unaltered habitat.

ACKNOWLEDGMENTS i

I am grateful to R. D. Ohmart for his advice and suggestions throughout the course of :
the investigation. R. J. Lederer and W. F. Laudenslayer provided useful comments on the '
manuscript. I thank the U.S. Forest Service, Tempe, Arizona, for their assistance, :
especially 1). R. Patton who offered advice and ideas. N. C. Schwertman kindly provided |
the statistical analyses. This study was supported by grants from the U.S. Forest Service j
(Cooperative Aid Agreements Nos. 16-382-CA and 16-402-CA for 1973 and 1974 respec- '|
lively). i

Franzreb • TREE SPECIES USED BY BIRDS

237

LITERATURE CITED

Anderson, R. 1960. Forest and shade tree entomology. John Wiley and Sons, New
York.

Austin, G. and E. Smith. 1972. Winter foraging ecology of mixed insectivorous bird
flocks in oak woodland in southern Arizona. Condor 74:17-24.

Balda, R. 1969. Foliage use by birds of the oak-juniper woodland and ponderosa pine
forest in southeastern Arizona. Condor 71:399-412.

Brewer, R. 1963. Ecological and reproductive relationships of Black-capped and
Carolina chickadees. Auk 80:9-47.

CoLQUHOUN, M. AND A. MoRLEY. 1943. Vertical zonation in woodland bird communi-
ties. J. Anim. Ecol. 12:75-81.

CoTTAM, G. AND J. CuRTis. 1956. The use of distance measures in phytosociological
sampling. Ecology 37:451-460.

Diamond, A. 1973. Habitats and feeding stations of St. Lucia forest birds. Ibis 115:
313-329.

Franzreb, K. 1975. The effects of timber harvesting on an avian community in a mixed-
coniferous forest. White Mountains, Arizona. Ph.D. thesis. Arizona State Univ.,
Tempe, Arizona.

. 1976. Possible avian habitat partitioning resulting from within-tree location

preferences in a mixed-coniferous forest. Sociobiology 2:103-106.

Gibb, J. 1954. Feeding ecology of tits, with notes on Treecreeper and Goldcrest. Ibis
96:513-543.

Hartley, P. 1953. An ecological study of the feeding habits of the English titmice.
J. Anim. Ecol. 22:261-288.

Hutchinson, G. 1958. Concluding remarks. Cold Springs Harbor Symposium. Quant.
Biol. 22:415-427.

Jackson, J. 1970. A quantitative study of the foraging ecology of Downy W'oodpeckers.
Ecology 51:318-323.

Kluijver, H. 1951. The population ecology of the Great Tit Parus m. major. L. Ardea
39:1-135.

Kocii, R., A. CouRCiiESNE, AND C. CoLLiNS. 1970. Sexual differences in foraging be-
havior of White-headed Woodpeckers. Bull. So. Calif. Acad. Sci. 69:60-64.

Levins, R. 1968. Evolution in changing environments. Princeton Univ. Press, Prince-
ton, New Jersey.

MacArthur, R. 1958. Population ecology of some warblers of northeastern coniferous
forests. Ecology 39:599-619.

. 1964. Environmental factors affecting bird species diversity. Am. Nat. 98:

387-397.

. 1968. The theory of the niche. Pp. 159-176, in Population biology and evolu-
tion ( R. Lewontin, ed.), Syracuse Univ. Press, Syracuse, New York.

AND J, MacArthur. 1961. On bird species diversity. Ecology 42:594-598.

Morse, D. 1967a. Competitive relationships between Parula W arblers and other species
during the breeding season. Auk 84:490-502.

. 1967b. Foraging relationships of Brown-headed Nuthatches and Pine W^arhlers.

Ecology 48:94-103.

. 1968. A quantitative study of foraging of male and female spruce-woods

warblers. Ecology 49:779-784.

Neyman, j. 1949. Contributions to the theory of x“ Pp. 239-273 in Proc. Berkeley
.Symposium on Mathematical Statistics and Probability. Univ. Calif. Press, Berkeley.

238

THE WILSON BULLETIN • VoL 90, No. 2, June 1978

Palmgren, P. 1930. Quantitative untersuchungen iiher die Vogelfauna in den Waldern
Siindfinniands. Acta ZooL Fenn. 7:1-218.

Shannon, C. 1948. A mathematical theory of communication. Bell System Tech. J.
27:379-423; 623-656.

Slobodkin, L. 1962. Growth and regeneration of animal populations. Holt, Rinehart,
and Winston, New York.

SouTiiwooD, T. 1961. The number of species of insect associated with various trees.
J. Anim. Ecol. 30:1-8.

Stallcup, P. 1968. Spatio-temporal relationships of nuthatches and woodpeckers in
ponderosa pine forests of Colorado. Ecology 49:831-843.

Sturman, W. 1968. The foraging ecology of Parus atricapillus and P. rujescens in the
breeding season, with comparisons with other species of Parus. Condor 70:309-322.

Williams, A. 1936. The composition and dynamics of a beech-maple climax com-
munity. Ecol. Monogr. 6:317-408.

Willson, M. 1970. Foraging behavior of some winter birds of deciduous woods. Con-
dor 72:169-174.

BUREAU OF LAND MANAGEMENT, 1695 SPRUCE STREET, RIVERSIDE, CA 92507.
ACCEPTED 15 JULY 1976.

Wilson Bull., 90(2), 1978, pp. 239-247

DOUBLE-BROODEDNESS IN PURPLE MARTINS IN TEXAS

Charles R. Brown

Most ornithologists consider Purple Martins {Progne subis ) to be single-
brooded throughout their range, although this is a somewhat controversial
topic. Allen and Nice (1952j, Lee (1968), Layton (1969), and Lowery
(1975) stated that martins are single-brooded with little or no evidence of
double-broodedness. Yet Forbush (1929), Baerg (1931), and Sprunt {in
Bent 1942) believed the birds to be double-brooded; Audubon (1840) men-
tioned third broods in Louisiana. Peterson (1941) and Harrison (1975)
indicated that martins occasionally raise second broods in the southern United
States. Johnston and Hardy (1962) recognized martins as usually single-
brooded but reported second broods in southern Illinois.

Allen and Nice (1952 ) questioned all second broods, providing testimonials
from southern martin fanciers to the effect that there is insufficient time
during the martin’s breeding season for double broods. Skepticism might
arise from a lack of details regarding reports of double-broodedness. Despite
the several references to second broods, the only detailed report in the litera-
ture is my ( 1973 ) account of a second brood in 1970 in north central Texas.

In 1976-1977 I observed 6 additional second broods by Purple Martins in
north central Texas. This paper will explain these occurrences, analyze en-
vironmental conditions and stimuli, and offer an hypothesis regarding double-
broodedness in Purple Martins.

DESCRIPTION OF THE MARTIN COLONY

The martin colony was located in my residential backyard within the city
of Sherman, Grayson County, north central Texas. In 1976-1977 the colony
contained 110 and 72 apartments respectively and consisted of 7 and 4 martin
houses of various sizes and designs. Fourteen other active martin colonies
were located within a 1.6 km radius from my colony. Prior to 1976 martins
had nested at this location lor 8 years.

In 1976, 35 pairs of martins attempted to nest in my colony. Of those,
18 (51.4%) appeared to be adult pairs and 17 (48.6%) appeared to be
subadult ( i.e., 1st year ) pairs. Sixteen (45.7% ) pairs raised 1 brood and
fledged young on the first attempt. Six (17.1%) pairs raised 1 brood and
fledged young on the second attempt, having failed for various reasons on
the first attempt. Nine (25.7% ) pairs failed to fledge young on either attempt.
Four (11.4%) pairs fledged second broods. In 1977, 27 pairs of martins
nested in the colony, but I did not collect detailed breeding data for that
year.

239

210

THE WILSON BULLETIN • Vol. 90, No. 2, June 1978

The 1976 figures were compiled after intensive surveillance of the martin
colony throughout the breeding season. I closely watched the colony for
several hours each day from the arrival of the first martins in February until
the last brood departed in August. I made nest checks daily. An assistant and
I banded 638 martins in the Sherman area in 1974^75. Several of these
banded martins were present at my colony in 1976.

OBSERVATIONS

1 spent 1000 (±5%) h watching the martin colony during the breeding
season of 1976. My intense surveillance of the colony was significant since
3 of the 4 pairs involved in second broods were unmarked. 1 spent many
hours watching these birds and am fully convinced of the validity of the
results. Individual recognition of unmarked birds is discussed further below.

Pair 1 : An all-purple adult male and an adult female arrived at the

colony on 20 February 1976. In contrast to many martins that claim several
rooms in the early part of the season, this pair claimed only 1 room through-
out the breeding season. Four young of their first brood left the nest on 29
May. After the young fledged, the adults led them back to the nest compart-
ment to roost nightly until 5 June.

On 5 June I began noticing Pair 1 perching by their nest and entering the
nest compartment. I assumed that the young had become independent and
that the adults were displaying post-breeding nest defense. Post-breeding nest
defense is common among martins. No juveniles returned to the nest on
5 June nor at any time afterward. Pair 1 remained at the colony for long
periods on 5-15 June but never brought nesting materials to the nest during
that time. On 16 June one egg was laid in the old nest, followed by another
on 17 June. The female began incubation on 17 June after laying the second
egg. The male was quite attentive to the female while she incubated, and
although male Purple Martins do not incubate, he frequently guarded the
nest when she was away. He also brought green leaves to the nest while the
female incubated.

The eggs hatched on 2 July. Both parents fed the young until they were
about 3 weeks old. At that time the male gradually stopped feeding. He
had ceased feeding completely by the time the young were ready to leave.
However, he often appeared and perched by the nest, and he and the female
were never hostile toward each other. When the young were about 2 weeks
old, the female ceased roosting in the nest and disappeared from the colony
each evening at dusk. Each morning she reappeared quite early. The male
had stopped roosting at the colony during the incubation period. One young
left the nest on 28 July, and the other left on 29 July. No juveniles returned to
roost in the nest after 29 July, and the parents rarely appeared.

Brown • DOUBLE-BROODEDNESS IN MARTINS

241

The unmarked male had no outstanding trait, but from watching his be-
havior very closely for many hours, I am convinced that the same bird was
engaged in both broods. The unmarked female was easy to recognize. She
was very white on the lower breast and belly, appearing much lighter than
other females that arrived in February. She also had a peculiar habit of run-
ning along the entire length of the porch of the martin house before flying.
She exhibited this peculiarity during both nestings.

Pair 2: Adult male arrived at colony on 24 February 1976 and began
claiming a tier of 4 nest compartments. An adult female arrived on 2 March
and established a pair bond with the male. Four young of a first brood fledged
from Pair 2’s nest on 31 May and 1 June. Soon after the young left, I
noticed another adult male claiming this tier of 4 rooms. When Pair 2
attempted to lead their young back to the martin house to roost on 1 June,
this new male attacked the juveniles and their male parent, but allowed the
female to land on the house and enter her nest. He would not allow the
juveniles or male parent to roost in the nest or anywhere else on the tier.
The female attempted to lead 1 or 2 of her young back nightly until 8 June,
but the new male never allowed the juveniles to roost. He tolerated the female
though and frequently courted and sang to her.

On 8 June the female began perching by the nest for long periods, and the
new male appeared to establish a pair bond with her. This “revised” Pair 2
spent much time sitting on the martin house on 8-12 June but never brought
nesting materials to the house. On 13 June an egg was laid in a room
adjacent to the former nest compartment. A small nest was in this new com-
partment, having been built by this same female earlier in the season when
she was building in all rooms on the tier. She laid 1 egg daily until 18 June
when she laid her 6th and final egg and began incubation. Her new mate
was delinquent in guarding the nest during the incubation period, rarely
appeared to guard in the afternoon, and rarely gathered green leaves.

The eggs hatched on 3 July. The female did considerably more feeding of
the young than did the male. Daily 2-hour watches of this pair revealed that
the female was responsible for 70-88% of the total feeding trips. However,
as the young became older the male appeared to develop stronger parental
instincts. When the young came out on the porch of the house 2 days prior
to leaving, the male guarded them virtually constantly while the female fed
them. Three young left on 29 July and 2 left on 30 July. One nestling had
disappeared from the nest a week earlier. The female roosted in a room of
the tier each night throughout the second nesting, but the male had ceased
roosting at the colony during the incubation period. The adults did not
lead the juveniles back to the nest to roost after 31 July.

Many times as I watched this new male of Pair 2, he sang to a female that

242

THE WILSON BULLETIN • Vol. 90, No. 2, June 1978

was feeding young on the tier below him. He often flew down to her and
entered her nest. I saw him feed this female’s young twice after finding his
young unreceptive, and he appeared to be a partial polygynist (See Brown
1975).

The original male of Pair 2 was easy to separate from the new male that
took over the nest after the first young left. The original male’s purple under
tail coverts were extensively edged with white. No other males in the colony
had such markings. From watching the unmarked female closely during both
broods, I am confident that she was the same individual involved in both.
During both broods, she was unusually aggressive toward a pair of House
Sparrows {Passer domesticus) that nested in a room on a lower tier. She
attacked the sparrows whenever she flew from the house. Also, while feeding
her second brood, she often first entered her nest of the first brood, then
“remembered” her second brood in the adjacent hole and moved to it.

Pair 3 : An adult male arrived on 18 March 1976 and established himself in
the attic hole of a small wooden martin house. He attracted an adult female
on 21 March. Three young of a first brood fledged from Pair 3’s
nest on 6 June. They brought their young back to the house to roost nightly
on 6-14 June. On 15 June 1 noticed Pair 3 perching by and entering their
nest. I often saw them near the nest on 15-19 June, but 1 never saw^ them
bring nesting materials to the nest during that time. On 16 June 1 saw
several males chasing the female of Pair 3 in a “rape” flight while her
mate tried to fight them off. 1 had never before seen a rape that late in the
season. Rape flights by martins are frequent in Sherman, Texas, in March,
April, and May while the birds are courting, but are very rare anytime in
June.

On 20 June 1 egg was laid in the nest, followed by 1 a day until 24 June
when the 5th and final egg was laid and the female began incubation. The
male was very attentive to his incubating mate and often guarded the nest
when she was away, but he rarely brought green leaves. The eggs hatched on
9 July. Both parents fed the young throughout the nestling period. Neither
adult fed at a greater rate than the other. Neither parent roosted at the
colony after the young were about 2 weeks old. Three young left the nest on
6 August, and 2 left on 7 August. The parents did not lead them back to the
nest to roost after leaving.

This male wore U.S. Fish and Wildlife Service band no. 772-05364. I con-
firmed this number with a 30X telescope and by capture during both nest-
ings. He had been banded as a nestling at a nearby colony in Sherman in
1974 and nested at my colony in 1975. The female was quite easy to distin-
guish from all other martins in the colony since the second, third, and fourth
primaries of her left wing were missing. The feathers were gone when she

Brown • DOUBLE-BROODEDNESS IN MARTINS

243

arrived in March, and no noticeable replacement occurred during the breed-
ing season. No other martin in the colony had such a noticeable “gap” in
its wing.

Pair 4\ An adult female arrived at the colony on 20 February 1976 and
began claiming a tier of 4 nesting compartments. An adult male, who was
paired to another female on an upper tier, showed interest in the female of
Pair 4 and claimed her until 6 March when another adult male arrived and es-
tablished a pair bond with her. Four young of their first brood fledged on 11
June. After the young fledged. Pair 4 led them back to the nest to roost
nightly until 20 June when I noticed the parents perching by the nest for
great periods. The juveniles did not return to the nest after 19 June.

Pair 4 spent much time sitting by, and in, the nest on 20-25 June, but never
brought nesting materials to the nest during that period. One egg was laid
in the nest on 26 June, followed by 1 a day until 29 June when the female laid
her 4th and final egg and began incubation. The male was quite attentive to
her while she incubated and often guarded the nest when she was away.
Virtually every morning until the eggs hatched, this male spent several hours
constantly gathering green leaves and bringing them to the nest.

The eggs hatched on 14 July. Both parents fed the young at about equal
rates until the young were about 3 weeks old. On 5-6 August the female
ceased feeding completely. During the last week before the young left, only the
male fed them, and the female only occasionally appeared and sat by the nest.
She and the male were never hostile toward each other. All 4 young fledged
from the nest on 11 August. The male did not lead them back to the nest to
roost after leaving. Both parents ceased roosting at the colony after the young
were about 2 weeks old.

The unmarked male had a throat mottled with purple and brown. He was
the only all-purple adult male martin 1 saw that was not uniform purple on the
throat. The unmarked female was an old appearing bird, and her breast and
belly were largely fuscous instead of dark gray. Her brownish breast con-
trasted sharply with her purple back. She also had an extremely vertical pos-
ture whenever she sat on the martin house. I am convinced that the same in-
dividuals were involved in both broods.

Colony Activity 13 June-11 August: From 13 June 1976 when the earliest
egg of a second brood was laid until 11 August when the young of the last
second brood departed, 21 martin pairs in the colony were engaged in various
nesting stages of first broods. The activity of these pairs is summarized in
Table 1.

Large numbers of martins visited the colony throughout 13 June-11 August.
I often saw 70 or more martins perched on martin houses and wires around
the colony while the second broods were in progress. Most of these were birds

244

THE WILSON BULLETIN • Vol. 90, No. 2, June 1978

Activity of Martins

Table 1
Raising First Broods at
IN Progress,

Colony

1976

While Second

Broods were

Nesting Stage

June 13

June 27

July 4

July 11

July 18

July 25

August 1

August 8

Feeding Young

Incubating

Laying

Yet to Lay

Figures indicate number of pairs and do not include 4 second broods.

that had completed breeding, but many were juveniles. Martins were present
during most of the day, with largest numbers appearing in the morning. I
had never seen such large numbers around the colony in mid- and late July
and early August. The many martins around the colony in July reminded
me greatly of a thriving colony in June. My studies indicate that many
martins becoming post-breeding vagrants and visit colonies in their local area
after finishing breeding. They show no definite breeding tendencies but sit :
on the wires and preen and alight on the martin houses. During Pair 4’s
last week, very few martins appeared at the colony in the mid-afternoon
hours, and Pair 4 reduced the frequency of feeding at that time.

DISCUSSION

I became fully convinced in 1976, through observations of 1 marked and

3 unmarked pairs, that Purple Martins occasionally raise second broods.

However, the color banded pairs in 1977 provide unequivocal evidence that i
Purple Martins are sometimes double-brooded. Since most field work was |
performed in 1976, the following discussion is based only on the 1976 double- j
brooded pairs and the breeding season of 1976. |

All 4 double-brooded pairs displayed normal breeding behavior while rais- I
ing second broods. I noted no appreciable behavioral differences among the |

4 pairs. They differed from first broods in only 2 aspects, both of which are j
probably insignificant. None of the pairs built a nest for their second brood, I
but all already had nests. Three pairs used the original nest hole for their '
second brood, while 1 used an adjacent compartment. None of the pairs l
led the young of their second brood back to the nest to roost for a long period :
after fledging. However, this was probably because the adults themselves |
were no longer roosting in the nest. I saw many martins of the neighborhood |

Brown • DOUBLE-BROOUEDNESS IN MARTINS

245

use a grove of eastern cottonwoods (Populus deltoides) for roosting each
evening in July and early August.

The male of Pair 1 and the female of Pair 4 ceased feeding the young of
their second brood before the young fledged. This loss of parental care may
be attributed to the lateness of the season, particularly in Pair 4’s case. Dur-
ing the last week when the female of Pair 4 had stopped feeding, I noticed
many male martins around the colony hut few females (often 25-35^ <5 ;19 ),
suggesting that females may move to the pre-migratory roosts earlier than
males.

Two of the double-brooded males often brought green leaves to the nest
during incubation. These double-brooded males gathered more leaves than
many single-brooded males, particularly single-brooded subadults.

All double-brooded martins observed were adults. Since adults arrive
on the breeding grounds before subadults, it is likely that all second broods
in martins are raised by adults. There is no time for later-arriving sub-
adults to rear second broods.

It is probable that all 4 double-brooded pairs successfully fledged their
first brood to independency. If the young of a first brood succumbed shortly
after leaving the nest, the parents’ subsequent second brood might be a re-nest.
This is unlikely in my cases though. All 4 pairs led their first brood back to
the nest to roost for a week or longer. It is likely that young martins become
independent 7-10 days after leaving the nest but I have not determined the
true time required for young to become independent.

Regarding Lowery’s statement (in Allen and Nice 1952) that late nesters
may occupy the nest sites after the original martins fledge young and can
be confused as second broods, I have found that even after fledging young,
many martins maintain close ties with the nest. Most pairs bring their young
back to the nest to roost nightly. Returning with the young at night serves
to discourage other martins from moving into the nest. Of my double-
brooded pairs, only the male of Pair 2 was displaced by an outsider who ap-
propriated the nest after the young fledged. 1 might also add that at my
colony in 1976 there were large numbers of vagrant martins that would have
likely moved in had the double-brooded martins not discouraged them by re-
turning each evening with their young. A further deterrent to outsiders
moving in is well-pronounced post-breeding nest defense in martins. Post-
breeding nest defense has been noted often in my studies and by Finlay
(1971).

I began sponsoring Purple Martin propagation efforts in the city of Sher-
man, Texas, in 1973-76 as an aid to my martin studies. Few martin houses
were present in the city in 1968 when I began studying martins and in 1970
when I noted an unsuccessful second brood attempt. By 1976 the number of

246

THE WILSON BULLETIN • Vol. 90, No. 2, June 1978

martin houses in the city had increased 550%. Populations of Purple Martins
increased similarly or more so.

Most martins in Sherman normally finish nesting activities in mid- to late
June. Until 1976 I had never recorded martins breeding at my colony after
10-15 July. A second brood in 1970 failed on 14 July. Few martins were
present around the colony then (Brown 1973). Numbers of post-breeding
vagrant martins that visit the colony also greatly decrease after 10-15 July.
Until 1976 I had rarely seen over 30 martins at the colony in late July and
August. I made a check of other colonies in Sherman in late July and found
few martins around any of them except mine and one small and newly-
established colony. An adult pair of martins at this small colony was feeding
young in early August. I suspect that this was also a second brood, but I have
no proof.

I was impressed by the many martins that raised first broods at my colony
in July (and 1 pair in August). More birds in 1976 led to many that were
unable to find breeding places until late May and early June. Consequently
these birds were still nesting in July. Purple Martin populations apparently
exploded in the Sherman area in 1976, resulting in many pairs nesting later
than usual. The presence of many martins late in the season undoubtedly
provided a stimulus for the rearing of second broods by pairs that fledged
young earlier. This is consistent with the second brood attempt in 1970
(Brown 1973) .

Johnston and Hardy (1962) believed that very mild and unseasonable
weather in southern Illinois provided stimulus lor second broods in that
area. But in Sherman, Texas, in 1976 weather conditions were normal, and
the first arrivals did not begin nesting any earlier than in past years. Pres-
ence of other martins late in the season apparently outranks weather condi-
tions as second brood stimulus for martins in north central Texas. Thus, I
suggest that second broods by Purple Martins may occur in the southern
U.S. whenever large numbers of martins are present at the colonies later
than usual. Such numbers may be due to either local population increases or
nesting disasters in the early part of the season.

Since most martins leave the Sherman area in mid- to late August, pre-
sumably on fall migration, juveniles of second broods have less time before
migrating to increase fat reserves and become proficient at catching insects.
This could result in greater mortality among juveniles of second broods which
would limit any inherited double-broodedness in a local Purple Martin popu-
lation.

SUMMARY

Four adult pairs of Purple Martins in a colony of 35 pairs in Sherman, Texas, success-
fully fledged second broods in the breeding season of 1976. Two additional pairs raised

Brown • DOUBLE-BROODEDNESS IN MARTINS

247

second broods in 1977. Breeding behavior of these second nestings was studied and was
found not to differ significantly from breeding behavior of first nestings.

Purple Martin propagation efforts in Sherman, Texas, in 1973-76 resulted in great
numbers of martins in the local area in 1976. Many martins were unable to find breeding
places in the early part of the season and raised their first broods much later in the
season than usual. Such late nesting apparently served as a stimulus to the raising of
second broods by pairs that fledged young earlier. Purple Martins may raise second
broods in the southern part of their range whenever large numbers of martins remain at
the colonies later than usual.

LITERATURE CITED

Allen, R. W. and M. M. Nice. 1952. A study of the breeding biology of the Purple
Martin (Progne subis) . Am. Midi. Nat. 47:606-665.

Audubon, J. J. 1840. The birds of America. Vol. 1. New York.

Baerg, W. j. 1931. Birds of Arkansas. Agr. Exp. Sta. Bull. 258.

Bent, A. C. 1942. Life histories of North American flycatchers, larks, swallows, and
their allies. U.S. Natl. Mus. Bull. 179.

Brown, C. 1973. A second brood attempt by the Purple Martin. Auk 90:442.

. 1975. Polygamy in the Purple Martin. Auk 92:602-604.

Finlay, J. C. 1971. Post-breeding nest cavity defense in Purple Martins. Condor
73:381-382.

Forbusii, E. H. 1929. Birds of Massachusetts and other New England states. Vol. 3.
Mass. Dept. Agri., Boston.

Harrison, H. H. 1975. A field guide to birds’ nests. Houghton Mifflin Co., Boston.
Johnston, R. F. and J. W. Hardy. 1962. Behavior of the Purple Martin. Wilson Bull.
74:243-262.

Layton, R. B. 1969. The Purple Martin. Nature Books Publishers, Jackson, Miss.
Lee, j. a. 1968. Relative reproductive efficiency of adult and subadult Purple Martins.
Chat 31:1-2.

Lowery, G. H., Jr. 1975. Louisiana Birds. 3rd ed. Louisiana State University Press,
Baton Rouge.

Peterson, R. T. 1941. The Purple Martin. Leaflet No. 13. Natl. Aud. Soc.

BOX 1309, AUSTIN COLLEGE, SHERMAN, TX 75090. ACCEPTED 3 DEC. 1976.

Wilson Bull., 90(2), 1978, pp. 248-260

FOOD OF NESTLING PURPLE MARTINS

Helene Walsh

Post-fleclging survival of young passerine birds appears to be strongly in-
fluenced by tbe quantity of food received while in the nest (Perrins 1965).
A number of studies have been done to determine how much nestlings are fed
but few (e.g., Royama 1966) have involved the greater part of the nestling
period.

Finlay (1971) used mechanically recorded visits to the nest cavity as
an index of feeding activity in Purple Martins {Progne subis) . He assumed
that food was brought to the nestlings on each visit. His results showed an
increase in visits to the nest cavity with increased brood size, but the increase
was not proportional to the number of nestlings involved. Finlay’s study
raises a number of questions concerning amount of food fed to individual
nestlings in broods of various sizes: (1) are parent birds able to maintain

constant energy intake by individual nestlings when brood sizes are larger by
increasing the proportion of feeding visits to the nest or, (2) can they com-
pensate for a larger brood by bringing either a greater biomass of food per
feeding visit or, (3) does the decreased surface-area-to-volume relationship in
larger broods enable individuals to maintain themselves sufficiently on a
lower biomass of food? 1 tried to answer these questions in Finlay’s study
area on the shore of Astotin Lake (53° 40' N, 112° 50' W) in Elk Island
National Park, Alberta, where martins were nesting colonially in artificial
nest boxes during the summers of 1970 and 1971.

MATERIALS AND METHODS

To answer the above questions I obtained qualitative and quantitative measurements of
food brought to nestling Purple Martins and analyzed these data with respect to brood
size, age of nestlings, time of day, and season. Unless otherwise stated the test for sig-
nificance used was Wilcoxon’s signed rank test (Steel and Torrie 1960) and the ac-
cepted level of significance was P < .05.

Nest boxes. — Purple Martins had inhabited the 4 wooden houses used during the study
for several years. I built an observation blind behind each house and had easy access
by moving a black cloth covering the back of each 18 cm” nesting compartment. A
periscope fastened over an opening in this cloth let me view activities without alarming
the birds.

Martin population. — Because the colony size was so small all nesting pairs were studied
in both years. In 1970 2 of the 4 pairs started laying about 1 week before the others.
Nestlings that hatched and survived were rearranged in these 2 nests to give broods of
3 and 2, to match those of the second 2 pairs of martins. In all nests some nestlings
died, probably a result of asynchronous hatching — the younger birds apparently could not
compete successfully for food. All the nestlings except for 2 in the older brood of 3

248

Walsh • FOOD OF NESTLING MARTINS

249

young died on the same day at the ages of 23 and 14 days, probably as a result of 3 con-
secutive cold rainy days.

In 1971 two broods were studied; 1 brood of 2 nestlings that hatched 8 July, and in
which 1 nestling died on day 19, and 1 brood of 4 nestlings, which hatched 14 July, and
in which 2 nestlings died on day 12. The above deaths were accidentally caused by a
food digestion study and resulted in brood sizes 2 and 1.

Food and food value. — The diet of nestlings was determined by collecting food given
them from days 2 to 23 post-hatching, using the method of placing a piece of pipe cleaner
around the neck of the nestling just tightly enough to prevent swallowing (Orians 1966).
After a feeding visit the nestling was immediately removed from the nest and the food
forced up to the beak by massaging the throat. Food gathered by this technique was
placed in a vial for subsequent weighing and identification. The volume of food col-
lected from each nestling was replaced by an approximately equal volume of food that I
caught.

Insects were identified to family with the aid of keys (Jaques 1947, 1951). A determina-
tion of the caloric value of the most common insects in the samples representing dif-
ferent families was done (Spice 1972) but results were inconclusive.

Food collection periods. — Collection of food samples at any one nest did not exceed 2
consecutive hours, with at least 4 h intervals between collections. Collection of food ter-
minated when nestlings were about 21 days old; at this age food began slipping past the
pipe cleaner, presumably a result of strengthened esophageal musculature. Nestlings
were weighed each day at about 18:00.

The birds’ activities at the nest were monitored by me over time periods varying from
1 to 3 h. Movements by adult martins in and out of the nest cavity were registered by a
mechanically activated event recorder.

W eather. — Data on meteorological conditions were collected at the colony site in both
years. Readings of light intensity, wind speed, temperature, and relative humidity were
made immediately before and after the observation and collection periods. General
weather data were obtained from the Edmonton International Airport, about 48 km from
the study area.

RESULTS AND DISCUSSION

Of 956 food samples collected from the nestling martins, 246 were collected
between 8 and 29 July 1970 from 4 broods, and 710 samples were collected
between 15 July and 11 August 1971 from 2 broods.

DIET OF NESTLINGS

Influence of time of year on diet.- — Families of insects contributing more
than 1% of the total weight of food collected in either year are listed in Table
1. Relative use of insect families differed between years; for the Nymphalidae
this difference was significant P 0.05) and appeared to reflect both
the 7 days’ difference in initiation of the nesting season, and the premature
deaths of the nestlings on 29 July 1970. In 1971 the study terminated on 11
August, 13 days later than in 1970. The mourning cloak butterfly {Nymphalis
antiopa) was the species most heavily taken, and as these do not start emerging

250

THK WILSON BULLETIN • Vol. 90, No. 2, June 1978

Frequency^ of

Insects ry

Table 1

Families Fed to Nestling Purple Martins at
National Park, Alberta

Elk Island

1970

1971

1970

1971

Family

(%)

Family

(%)

Aeschnidae

Coenagrionidae

Syrphidae

Muscidae

Nymphalidae

Apidae

Chironomidae

Cicadellidae

Siricidae

Order: Trichoptera

Formicidae

Corixidae

Cerambycidae

Others

Li})ellulidae

1 Based on percentage of total weight collected ( 1970, 101.2 g wet weight; 1971, 278.7 g wet
weight).

as adults much before the first of August (Can. Dept. Agric. 1958) they were
not available for the nestlings studied in 1970. The insignificant difference in
biomass of aeschnids (dragonflies) taken is not unexpected because pre-
mature death of the nestlings in 1970 biased the food samples towards those
families containing small insects that are fed to younger nestlings. The dif-
ference in the use of syrphids (flower flies) between the 2 years was signifi-
cant (x“, P < 0.05). This seemed to be because syrphid activity is greatly
affected by weather conditions. (Under sunny conditions they contributed
24% of diet compared to 3% under cloudy, Spice 1972.) Data from the
Edmonton International Airport indicated that 1971 was sunnier (mean hours
of sunshine per day 8.7 in 1970 and 12.1 in 1971), drier (total precipitation
11.2 cm in 1970 and 3.1 cm in 1971), and warmer (mean temperature 21.7 C
in 1970 and 25.0 C in 1971), so based on weather conditions this difference
in use of syrphids between years is not unexpected.

Influence of time of day on diet. — The martins’ use of various families of
insects depended on the time of day the birds were feeding. Of the 4 insect
families taken most frequently, aeschnids were captured throughout the day,
chironomids (midges) only in the morning, and nymphalids (butterflies)
and syrphids mainly around midday ( Spice 1972) .

To investigate the relationship between time of day and amount of food
supplied to nestlings, the rate at which food was supplied was estimated by
averaging the weight of food brought per hour for each hour of the feeding
period (Fig. 1). I feel that the fluctuation in amount of food brought to the
nestlings was a reflection of begging intensity of nestlings, which seemed to
influence the number of feeding trips by adults. Begging intensity was higher

Walsh • FOOD OF NESTLING MARTINS

251

Fig. 1. Rate of feeding of nestling Purple Martins expressed as mean biomass (g dry
weight) fed per hour. Data from 1971. Hours involved indicate the hours the author
spent collecting food from nestlings at that time of day in 1971.

in the morning, declined after a feeding period, and cycled in this manner
throughout the day. The degree of fluctuation may reflect both the avail-
ability of food and the hunger of the adults themselves.

Influence of nestling age on diet.— The relative use of various insect
families over the nestling period in 1971 is shown in Fig. 2. The size of the
insects used in these families ( along with their availability ) could account for
changes in use with age: syrphids <2 cm long, aeschnids and nymphalids <3
cm long.

Role of parents in supplying food. — As nestling age increased, parental
feeding increased, and for awhile most of this activity was performed by the
male. However, as the need to brood the young decreased, the female pro-
vided more food and the male’s proportion of feedings was reduced corre-
spondingly (Spice 1972).

EFFECTS OF BROOD SIZE

Visits to nest. — A nonproportional increase of adult visits to the nest with
increasing brood size has lieen generally found (Moreau 1939, 1947; Lack

252

THE WILSON BULLETIN • Vol. 90, No. 2, June 1978

TOTAL FOOD
(G DRY WT)
70

4.4 4.0 5.2

■ AESCHNIDAE
B CHIRONOMIDAE
B NYMPHALIDAE
H SYRPHIDAE
□ OTHERS

TOTAL FOOD
(G DRY WT)
60'

4.7

o 50-

5.8 3.7 3.9 3.7 8.6

5.0 7.0 5.2

6.0 3.2

AGE OF NESTLINGS (DAYS POST-HATCH)

Fig. 2. Use of insect families in relation to age of nestling Purple Martins expressed
as % of total food collected for that age. Only those families contributing 10% or more
of the weight of food gathered at a specific age are figured. Key to letters above bars:
b, Asilidae; d, Coenagrionidae; f, Formicidae; g, Libellulidae; k, Siricidae.

Walsh • FOOD OF NESTLING MARTINS

253

NUMBER OF

SAMPLES 1970 20 13 9 19 is le 25 17 22

AGE OF NESTLINGS (DAYS POST-HATCH)

Fig. 3. Mean weight of food samples brought to nestlings by adults throughout the
nestling period. Mean brood size in 1970 = 2.5; in 1971 = 3.0.

and Silva 1949; Kendeigh 1952; Gibb 1950, 1955; Morehouse and Brewer
1968; Finlay 1971). However, the assumption that nestlings belonging to
larger broods receive less food does not necessarily follow. Adults having
larger broods could compensate in several ways to meet their nestlings’ energy
requirements.

Biomass of food. — One way to compensate would be for the parents of
larger hroods to bring food samples of relatively greater biomass. To invest-
igate this, the mean weight of each food sample brought per visit was com-
pared with increasing age of nestlings (Fig. 3). In both years mean weight
of food brought per feeding trip increased with age; thus the adults were
compensating for size of nestlings. However, the data do not indicate a com-
pensation for brood size. The mean values for 1970 were significantly higher
than in 1971, but the average brood size in 1970 was 2.5 nestlings compared

254 the WILSON BULLETIN • Val. 90, No. 2, June 1978

AGE OF NESTLINGS (DAYS POST- HATCH)

Fig. 4. Percentage of nest visits in which adults brought food. Curve fitted by in-
spection to data from 1970 and 1971. “Small” brood size refers to a brood of 2 nestlings
reduced to 1, 19 days post-hatching. “Large” refers to a brood of 4 nestlings reduced to
2, 12 days post-hatching.

to 3.0 in 1971. If the adults were compensating for brood size, food samples
should have been heavier in 1971 than in 1970. The difference in mean weight
of food samples between the years probably reflected weather conditions, which
in 1970 were cooler and cloudier. This was correlated with a greater diversity
of prey items in the diet, the majority of these also being smaller items. Two
possibilities exist: smaller prey items are coated with more saliva by the
adult, thereby increasing the average wet weight recorded per collected sample ;
or, the adults were collecting a greater total weight of food material per feed-
ing trip when prey items were small in size.

Brood size related to weight of food brought per feeding trip was also in-
vestigated using the 1971 data. There was no significant difference in the
weights of food samples brought to the 2 broods. Royama’s (1966) results
differed from mine in that he found average weight of prey brought per feed-
ing trip was heavier among tits feeding smaller broods. This he explained by
saying that begging intensity in smaller broods was less and thus gave adults
time to select larger prey items. This difference between species may reflect
the fact that whereas tits usually bring but one food item per trip, martins
frequently bring more than one.

Walsh • FOOL) OF NESTLING MARTINS

255

HRS OBS
SMALL
LARGE

8 5.0

2 4.0H

CO
LU

z 3.0H

2.01

S 1.0H

6,5 12.5 10.7 85 12.3 120 90 70 6 3 5.5 43 8.3 4.0 7.3 7.5 5.0 4.5 5.0 4.5 7.0 4.0 2.5 1.5
4.0 4.0 30 8.5 5,0 4.5 6.8 5.5 4.5 7.8 6.3 3.8 3.0 4.3 5.8 5.0 7.5 5.8 83 3.3 8.0 8.8

• SMALL *

X LARGE •

• X

• •

I X

X X

I I I r I I I I I I I I I I I I I I I I I

5 10 15 20

AGE OF NESTLINGS (DAYS POST-HATCH)

Fig. 5. The effect of brood size on the hourly number of food-bearing visits per nest-
ling Purple Martin. Data from 1971. (See Fig. 4 for explanation of “small” and “large”
broods.)

Frequency of feeding trips. — Adults could also compensate for a larger
brood by increasing feeding frequency. Figure 4 shows that the percentage of
food bearing visits, regardless of brood size, increased with the nestlings’ age.
Figure 4 also reveals that adults feeding larger broods carried food on a sig-
nificantly greater percentage of trips to the nest. This factor has not been
considered by many workers who have assumed that all trips to the nest cavity
by adults are equal to, or proportionately equal to, the number of food hear-
ing trips.

To determine if this method of compensation gave an individual nestling
of the larger brood the same number of actual feedings as those given an
individual of the smaller brood, the number of times nestlings were fed per
hour was investigated in the 2 broods of different size in 1971 (Fig. 5). Nest-
lings in the smaller brood were fed significantly more often than were those
in the larger brood. Similar results were seen in the data from 1970, hut
sample size was such that values could only he determined up to day 12.

256

THE WILSON BULLETIN • Vol. 90, No. 2, June 1978

20-

-- 15

10*

O 3 NESTLINGS 1970
□ 2 NESTLINGS 1970
A 2 NESTLINGS 1971

s ° .

I I

2 9

A O ^

I I I I I I I I I I I ■ I ■ r

5 10 15

AGE OF NESTLINGS (DAYS POST-HATCH)

Fig. 6. Effect of brood size on calculated total daily weight of food per nestling.
Open symbols used for data collected under good weather conditions, and closed symbols
for data collected under cold or rainy conditions.

Daily food intake. — Knowing the mean weight of food brought to the nest
per trip and the number of such trips per unit of time for broods of different
sizes, it was possible to calculate the total food intake for each brood ( Fig.
6 j . The total weight of food supplied in 1970 was significantly higher than in
1971, probably because of weather conditions, which in 1970 were generally
colder than in 1971, and presumably the 1970 nestlings needed more food.

Apparently, food intake of nestlings, per unit time, increased with age to
about day 13, and then leveled out ( Fig. 6 ) . Results beyond day 18 were not
obtained because at this age nestlings began coming to the door to receive food
and the adults were able to feed them without activating the recorder.

Nestling growth curves. — Theoretically, nestlings from smaller broods,
which received more food per unit of time than did nestlings from larger
broods, should exhibit correspondingly different growth curves (Fig. 7).
Such seemed to be the case as shown by the data for 1970 and 1971. In each
case the smaller brood attained the greatest average individual weight. The
difference was not apparent early in nestling life in 1970. This can probably
be explained through differential heat loss in different-sized broods. This loss
among nestlings of larger broods should be less because of the surface-area-

U'alsh • FOOD OF NESTLING MARTINS

257

Fig. 7. Effect of brood size on growth curves. In 1970 broods on the same graph
hatched the same day. In 1971 the “small” brood hatched 6 days later than the “large.”
“Small” broods had 2 nestlings and “large” had 3.

to-volume relationship, and thus until homeothermy is attained, a larger brood
size would be advantageous during a colder season, such as 1970.

Brooding activity. — It has been suggested that, because the surface-area-to-
volume ratio is greater for smaller broods, individuals in small broods lose
more heat and therefore require more food to maintain themselves than do
nestlings of a larger brood size ( Royama 1966 ). It seems possible that in-
creased brooding activity of adults could compensate for this to some extent.
Amount of time spent not brooding by females with small and large broods
was analyzed for 1971 (Fig. 8). The percentage of time nestlings were
brooded decreased with age to day 16 or 17, after which females ceased brood-
ing. Slopes of regression lines were similar for both brood sizes, however,
percent of time spent brooding the larger brood was significantly less than
that spent brooding the smaller. Royama ( 1966 ) found a similar situation
among Great Tits [Parus major). This suggests that the parents do reduce
heat loss by their nestlings through brooding behavior. This observation has
general implications for clutch size. Adults with a small brood can more

258

THE WILSON BULLETIN • Vol. 90, No. 2, June 1978

Fig. 8. Brood size related to percentage of time nestling Purple Martins were not
brooded throughout the nestling period. Data from 1971. Regression lines: Large, T =
8.9 + 5.8X; Small, Y = -10.1 + 6.1X.

easily provide them with food and intensive brooding. As brood size increases,
and the surface-area-to-volume relationship decreases, heat loss by nestlings
decreases. Thus, although adults must spend more time feeding young in
large broods they can accomplish this by spending less time brooding. Hence,
for any particular year there is an optimum brood size that permits adults to
feed and brood the young with greatest success; beyond this brood size sur-

Walsh • FOOD OF NESTLING MARTINS

259

vival of nestlings declines. It is this optimum brood size, which varies from
year to year depending on weather conditions that probably determines what
the mean clutch size will be for an area.

SUMMARY

A study of the food of nestling Purple Martins at Elk Island National Park, Alberta,
during the summers of 1970 and 1971 indicated that insects of the families Syrphidae,
Nymphalidae, and Aeschnidae comprised the majority of their diet. The quality and
quantity of food fed to the nestlings changed with season, time of day, and age of
nestlings.

The influence of brood size was investigated to see if adults were compensating for the
energy requirements of a larger brood in ways other than increasing the number of visits
to the nest. They did not bring a greater biomass of food per feeding trip. They did
increase the percentage of trips to the nest that were feeding trips, hut, in spite of this,
individual nestlings from smaller broods were fed more often than those from larger
broods. This was reflected in nestling growth curves in which nestlings of smaller broods
attained a greater weight prior to fledging. Adults with larger broods spent less time
brooding nestlings than did adults with smaller broods, possibly a result of a smaller
surface-area-to-volume ratio, and consequently lesser heat loss by nestlings in larger
broods.

ACKNOWLEDGMENTS

Support for this study was provided from the Lniversity of Alberta, and from grants
to myself and to D. A. Boag from the National Research Council of Canada.

I wish to thank D. A. Boag, B. Hocking, V. Lewin, J. C. Finlay, D. Griffiths, R. Walsh,
and the administrative staff of Elk Island National Park for their help and cooperation
during the course of the study. Thanks also to O. L. Austin, R. F. Johnston, H. W. Kale,
J. Jackson, A. S. Gaunt, M. McNicholl, and K. Martin for assistance with the manu-
script.

LITERATURE CITED

Canadian Dept. Agric. 19.58. Forest Lepidoptera of Canada. Puhl. 10.34, vol. 1.

Finlay, J. C. 1971. Breeding biology of Purple Martins at the northern limit of their
range. Wilson Bull. 83:255-269.

Gibb. J. A. 1950. The breeding biology of the Great and Blue titmice. Ibis 92:507-
539.

— . 1955. Feeding rates of Great Tits. Br. Birds 48:49-58.

Jaqles, H. E. 1947. How to know the insects. Wm. C. Brown Co., Dubuque, Iowa.

. 1951. How to know the beetles. Wm. C. Brown Co., Dubuque, Iowa.

Kendeigii, S. C. 1952. Parental care and its evolution in birds. Univ. Illinois Press.
Lrhana.

Lack, D., and E. T. Silva. 1949. The weight of nestling Robins. Ibis 91:64-78.

Moreau, R. E. 1939. Numercial data on African birds’ behavior at the nest: Hirundo
s. smithii Leach, the Wire-tailed Swallow. Proc. Zool. Soc. London 109A :109-125.

. 1947. Relations between number in brood, feeding rate and nestling period in

nine species of birds in Tanganyika Territory. J. Anim. Ecol. 16:205-209.

260

THE WILSON BULLETIN • VoL 90, No. 2, June 1978

Morehouse, E. L. and R. Brewer. 1968. Feeding of nestling and fledgling Eastern
Kingbirds. Auk 85:44-54.

Orians, G. a. 1966. Food of nestling Yellow-headed Blackbirds, Caribou Parklands,
B. C. Condor 69:321-337.

Perrins, C. M. 1965. Population fluctuations and clutch size in the Great Tit, Parus
major L. J. Anim. Ecol. 34:601-647.

Royama, T. 1966. Factors governing feeding rate, food requirement and brood size
of nestling Great Tits i Pams major) . Ibis 108:313-347.

Spice, H. 1972. Food habits of nestling Purple Martins. MSc. thesis, Univ. Alberta,
Edmonton.

Steel, R. G. D., and J. H. Torrie. 1960. Principles and procedures of statistics.
Toronto, McGraw-Hill Book Co., Inc.

DEPT. OF ZOOLOGY, UNIV. OF ALBERTA, EDMONTON, ALBERTA, CANADA. PRESENT
ADDRESS: GENERAL DELIVERY, GROUARD, ALBERTA, CANADA. ACCEPTED 10

FEB. 1977.

Wilson Bull, 90(2), 1978, pp. 261-270

REPRODUCTION AND NEST SITE SELECTION BY
RED-WINGED BLACKBIRDS IN NORTH LOUISIANA

Bryan T. Brown and John W. Goertz

The purposes of this study were to investigate reproductive variations
and nest site selection of Red-winged Blackbirds {Agelaius phoeniceus) in
primarily upland habitat in north Louisiana. This study provides further
information on the nesting ecology of the Red-wing and research that is
necessary in compiling an accurate picture of Red-wing population ecology
throughout North America, as stressed by Gottschalk ( 1967 ) . In addition,
knowledge of local breeding populations is important, since most blackbird
damage to rice is done by resident birds < Meanley 1971).

STUDY AREA AND METHODS

Most nests were found in Lincoln Parish, Louisiana, although significant numbers were
found in other portions of north Louisiana within an area delimited by the borders of
Texas and Arkansas, the Ouachita River, and the city of Alexandria. The study area, re-
ferred to as the Northwest Louisiana Uplands fSt. Amant 1959), is generally hilly with
elevations ranging from 15 to 161 m above sea level. A few rivers traverse the area, hut
most streams are small and slow-moving, subject to rapid rises and flooding. After re-
forestation, much of the area has developed a heavy second growth of native pines and
hardwoods. The clearing of these forested areas along streams to provide pastures for
cattle, man-made reservoirs, and small farm ponds has increased available Red-wing
nesting habitats.

A total of 755 Louisiana Tech University museum record cards for active nests from
1963 to 1975 were available for analysis. Although some nests, represented by nest cards,
were originally located by undergraduate ornithology students, all nests were checked

(authenticated) one or more times by at least 1 of us or by ornithology graduate stu-

dents who had previously conducted nest studies. In some instances observations
denoted on the cards were incomplete. This is, in part, the reason for the variations
in numerical totals for different categories of data.

During each of the annual nesting seasons a rather equal amount of time and effort
was allotted to an overall search for nests, eggs, young, and related data. However,
some types of data were collected only during 1975; these are so indicated when presented.

Various clues suggested the predators responsilde for a nest loss. Eggshells present
around the nest site indicated a mammalian predator. Loss of 1 or 2 eggs possibly in-
dicated an avian predator, as did the presence of “peck holes.” Snakes and mammals
usually ate all eggs in a nest. Robertson (1972) stated that egg loss in nests not ac-
companied by significant nest damage was due to snakes or birds.

The season was divided into 2-week periods for the purpose of data analysis. Nests
containing completed clutches that were discovered during the first 5 days of a 2-week
period were considered to have been active during the immediately preceding period;
as were nests with young. Those nests found that contained eggs during the first 5 days
of a 2-week period were assumed to have been built in the preceding period; nests that
contained young on the first 5 days of a period were assumed to have been built 2 periods

261

262

THE WILSON BULLETIN • Vol. 90, No. 2, June 1978

Table 1

Chronology of Nesting as Indicated
Eggs, and Young Present

BY THE Observed
IN 2-week Periods

Number of Active
, 1963 TO 1975

Nests,

April

May

June

July

August

Category 1-15 16-30

1-15

16-31

1-15

16-30

1-15

16-31

1-15

Number of

active nests 2 102

535

243

102

Number of

eggs 7 181

1405

522

181

148

Numlier of young 4

303

211

previously. Nest building time was assumed to take 6 days (Bent 1958). Young per
nest calculations are based on nests containing young, and not on total nests.

Only those eggs actually observed were included in the total number of eggs, including
known unhatched eggs, but eggs as indicated by eggshells were not included. A com-
plete clutch was one in which 2, 3, 4, or 5 eggs were successively counted during suc-
cessive observations. Certain eggs were measured at random throughout the 1975 season
with a Vernier beam caliper. If both eggs and young were present, the date the eggs
were laid was calculated by adding the age of the young ( Bent 1958, Holcomb and
Tweist 1971) to the incubation period of 10 to 12 days (Allen 1914). The number of
eggs and young present during 2-week periods was determined by tallying only the
number of eggs of young observed during the period.

RESULTS AND DISCUSSION

Active nests with eggs were present from 7 April to 2 August (Table 1).
Similarly, they occurred from 7 April to 23 August in Arkansas (Meanley
1971) and 30 April to 6 August in Ohio ( Dolbeer 1976). The largest num-
ber of active nests occurred during the first half of May as were those
recorded by Smith (1943) in Illinois. Elsewhere, the height of the nesting
season occurred later: mid May in Ohio (Dolbeer 1976), late May in

Connecticut (Robertson 1973) and New York (Allen 1914, Case and Hewitt
1963 ) among marsh nesters, but seems to occur in early June among upland
nesters in New York.

Inactive nests were not tallied prior to 1975, but in 1975, 24 of 151 nests
(16% ) followed from the time nests were built were found in which no eggs
or young were known to have been present; some of these may have been
victims of snake predation. Bent (1958) and Goddard and Board (1967)
found that many completed Red-wing nests were never used.

Of 2178 eggs recorded, 1461 made up 382 known completed clutches, for
an average of 3.82 (range, 2-5) eggs per clutch. Two completed clutches
had 2 eggs; 73, 3 eggs; 297, 4 eggs; and 10, 5 eggs. This was higher than

Brown and Goertz • RED-WINGEU BLACKBIRDS IN LOUISIANA 263

the average number of eggs reported by Robertson (1973) in Connecticut
(3.37 to 3.50); by Dolbeer (1976) in Ohio (3.38 j; by Beer and Tibbitts
(1950) in Wisconsin (3.7); by Orians (1961) in California (3.45 to 3.75) ;
or by Meanley (1971) on the Arkansas Grand Prairie ( 3.2). Incubation
usually begins after the third egg is laid (Bent 1958). But, in this study,
many nests containing only 3 eggs were not included as completed clutches
due to insufficient observations. If all clutches containing 3 or more eggs are
included as completed clutches, then, average clutch size for this study is
3.66.

Bent ( 1958 ) reported that the range of eggs per completed clutch was
3 to 5. Goddard and Board (1967) in Oklahoma, and Case and Hewitt
(1963) in New York found 1 to 5 eggs per clutch, while Orians (1961) in
California found 1 to 6, and in this study, as well as Dolbeer’s (1976), the
range was 2 to 5. We found 2 nests in which there were 2 known eggs, laid,
incubated, hatched and young successfully fledged. It is possible that some
eggs were removed from these nests by predators or cowbirds or accidentally
ejected. Since the active periods for these 2 nests were late June and early
July it is also possible that these were second broods or renesting attempts.
Double brooding (Meanley 1971), or renesting attempts is common (Smith
1943, Bent 1958, Goddard and Board 1967, Dolbeer 1976), Goddard and
Board (1967) reported that clutch size does decrease later in the nesting
season and Dolbeer (1976) found 2-egg clutches common late in the season
in Ohio.

During the earliest and latest parts of the nesting season, clutches of 3 eggs
were most common, whereas clutches of 4 eggs were most common during the
peak of the nesting season (Table 2). Clutches containing 5 eggs were pres-
ent only during the height of the nesting season.

The incubation period was 11 to 13 days. There were 36 known unhatched
eggs among incubated clutches, 25 of which were recorded in 1975 when
careful observations were made for this detail.

Average egg size of 303 eggs was 17.4 X 23.6 mm. Egg width ranged
from 15.9 to 18.8 mm, whereas length ranged from 21.0 to 28.1 mm. Egg
size was similar to the average of 17.5 X 24.8 mm for 380 eggs in the United
States National Museum (Bent 1958). The lower range in both egg length
and width is approximately the size of eggs laid by the Brown-headed Cow-
bird (this study and Bent 1958). There were no notable variations in egg
size as the season progressed. Statistical analysis of average egg sizes sea-
sonally showed no significant difference to exist ( t-test. P>.05).

Nests were found in 30 species of plants, as compared to 30 in New York
(Case and Hewitt 1963); 30 in Elorida (Stowers et al. 1968): and 70 in
Arkansas ( ^leanley 1971). Major Plant types used for nest support were
buttonbush (Cephalanthns occidentalis) . willow (Salix spp.), rush (Juncus

264

THE WILSON BULLETIN • VoL 90, No. 2, June 1978

Table

Maximum Number

OF Eggs

Known per Nest During

2-week

Periods, 1963 to 1975

Number of April

May

June

July

August

nest a

a b

39 24

55 18

3 1

119 56

4 1

214 71

7 2

Average

number of

eggs per

nest 3.5

3.0

3.2 3.1

2.8

2.6

2.7

3.0 2.5

spp.) bulrush {Scirpus

cyperinus ) , cattail {Typha spp.),

common alder

{Alnus serrulata)

, sweetgum { Liquidambar styraciflua) .

, dock

{Runiex spp.j

and grasses (Grarjiinae)

(Table 3). Plants of minor importance (used 1 to 9

times ) were pine

( Pinus

spp. ) , cypress (

xodiuni sp.),

silverling (

Baccharis

halimifolia) , vasey-grass

. {Paspalum urvillei).

privet 1

1 Ligustrum

sinense ) ,

wax-myrtle i Myrica cerifera ) , peach

( Prunus persica ) ,

horse-chestnut

(Aesculus pavia)

, oak (

Quercus spp. ) ,

live-oak

( Quercus virginiana ) , com-

mon elder {Samhiicus

canadensis ) , ironweed

( y ernonia ^ :

sp. ) ,

hawthorn

{Crataegus spp. j

, redbud {Cercis canadensis).

rose (

Rosa

sp. ) ,

sassafras

{Sassafras albidujji), blackberry {Rubus sp. ) , sumac {Rhus

sp. ) ,

thorough-

Table

Season Use of

Plant

Species for Nest Support

BY Red-winged

Blackbirds*

April

May

June

July

Supportive

plant a

Buttonbush **

104

W411ow 1

Rush

Bulrush

Cattail 1

Common Alder

Sweet gum

Dock

Grasses 1

* Includes 498 nests, 1964-1975.

** Nests not observed but probably did occur.

Brown and Goertz • REU-WINGEI) BLACKBIRDS IN LOUISIANA 265

Heights of 393 Red-winged

Table 4
Blackbird Nests,

Listed by

Habitat Type'

Habitat

No. of plant

species used No. of nests

Ave. height of
nests in m
( range )

FRESHWATER AREAS
Seasonally flooded
basins ( ditches, etc. )

117

1.2 (0.3-3.1)

Open freshwater
areas (edge)

173

1.0 (0.2 4.0)

UPLAND AREAS
Pasture (edge)

1.4 (0.3-7.6)

Cultivated areas

(orchards, arboretum, etc.)

1.1 (0.2-3.0)

TOTALS

393

1.1 (0.2-7.f3)

1 Habitat t>i)es from Shaw and Fredine (1956).

wort {Eupatorium perfoliatum ) , persimmon { Diospyros virginiana) and
cornel iCornus foemina) .

Of 554 nest sites for which the plant species was known, 261 (47%) were
in buttonbush, a common hydrophilic shrub of marshes, lakes, and low
pastures throughout the Southeast. Stowers et al. ( 1968 ) found buttonbush
to comprise 28% of the plants used for nesting by the Red-wing in Florida.
Cattail (90%) and willow (4%) comprised the majority of supportive plants
used in Oklahoma (Goddard and Board 1967), whereas Robertson (1972)
found cattail and dock to be the most used plants in Connecticut.

Variation in the average height of Red-wing nests within different habitats
is shown in Table 4. The ratios of plants used changed from one habitat to
another, as did plant occurrence ( Table 4 ) . In addition, the height of nests oc-
curring in single plant species may differ with a change in habitat. On
upland sites, it appeared to Francis (1973) that nest site selection was in-
dependent of height above the ground, but was related to the distance below
the top of the vegetative canopy. Heights may he associated with the chang-
ing growth form of plants dictated by different environmental conditions in
each habitat. Most nests tended to be lower when over open water, and
higher when over seasonally flooded basins or pastureland (Table 4), al-
though this was not always true. The largest number of nests occurred in
open freshwater areas; possibly indicating that open freshwater is the pre-
ferred habitat in north Louisiana. A pooled t-test analysis of nesting habitats
and nest heights (Table 4) indicates that a very highly significant difference
in nest height existed between the 2 freshwater areas (P < .001). In addition.

266

THE W ILSON BULLETIN • Vol. 90, No. 2, June 1978

Table 5

Nest Height and Nesting Success for 126 Active Nests, 1975

Height (m)

Number
of nests

Number

successful

Percent

successful

0.0-0.6

12.5

0.6-1.2

28.1

1.2-1.8

32.5

1.8-2.4

33.3

2.4-8.0

21.5

Totals

126

27.8

a significant difference (P<.05) existed between upland and freshwater
areas. Lesser differences were noted between upland habitats iP <0.10j.

Nesting success was compared with nest height (Table 5). depth of water
below the nest (Table 6 1, and among the 6 major supportive plant types
(Table 7 ) .

From 1963 to 1975, 65 known successful nests were recorded. A more
valid comparison of successful nests is given by careful observations made
during 1975: 35 known successful (23.1%) nests and 9 probably successful
(5.9%) out of 151. However, if only active nests are compared, 44 of 127
(35%) were successful, compared with 43% reported by Bent (1958),
29% in \^isconsin (Young 1963), 32% in New York (Case and Hewitt

Depth

OF Water

Table 6

Below Nests and Nesting Success

Depth of water

Number

Percent

under nest ( cm )

of nests

successful

127 ACTIVE NESTS ( 1975)

0 76

35.1

1-15

17.9

over 15

33.3

totals

127

27.6

ISOLATED BREEDING

POPULATION ( 1975)

4-1.0

1-15

33.0

over 15

50.0

totals

43.0

Brown and Goertz • RED-WINGED BLACKBIRDS IN LOUISIANA 267

1963 j, 50% in Maryland ( Frankhauser 1964 j, and 53% in Pennsylvania
(Brenner 1966 j. Nesting success of this study may be small due to individual
breeding populations covered in this analysis having been relatively small
compared to larger breeding populations that are often reported in extensive
marshes. In larger nesting colonies, predation pressure per individual nest
is much lower than in small colonies ( Darling 1938, Fautin 1941, Smith 1943,
Robertson 1973 j. However, in this study, an upland breeding population
of 42 nests had 43% success in 1975, a slightly higher success rate than all
other nests combined for that year (Table 6).

We observed 685 young in 239 nests for the average of 2.87 (range, 1-5)
young per nest. An average of 2.77 young per nest were fledged from the
65 successful nests, from 1963 to 1975. For the 1975 season alone, 2.97 young
were known fledged per successful nest. The average number of young
fledged per active nest (N = 127) in 1975 was 0.83. This compares to 0.6
in New York (Case and Hewitt 1963), and as high as 1.9 in Ohio (Francis
1975 ) . Dolbeer ( 1976 ) reported the average number of young fledged
annually per female as 1.3 in Ohio. In this study, young were present from
25 April to 23 July, with the largest number of young recorded in early May
(Table 1).

Meanley (1971) reported that nesting success was higher in nests more
than 0.6 m above the ground, while Goddard and Board (1967) noted that
lower nests were more successful; however, the majority of those nests were
over water in cattails. Data in Table 5 indicate that nests in the 1.2 to 1.8 m
range and 1.8 to 2.4 m range were similarly successful (32.5% and 33.3%)
and nests at 0.6 to 1.2 m were only slightly less successful (28.1%). Low
success (12.5%) for nests 0 to 0.6 m high was due, largely, to predation.
The highest nests (over 2.4 m) were not necessarily the most successful
(21.5%). This may have been due to weather factors, e.g., those nests
knocked down by wind and rain ( Francis 1971). Goddard and Board (1967 )
and Robertson ( 1972 ) observed that nesting success was greater as the
depth of the water below the nest increased. The smaller number of nests
over deep water were as successful (33.3%) as all the combined nests over
land (31.5%) (Table 6). The nests over shallow water (1 to 15 cm) had poor
nesting success (17.9%). This may be due, in part, to the fact that nests
over water tend to be lower (Table 4), whereas the shallow water would not
afford the protection of nests over deeper water (Table 6).

Nesting success is also partially related to sturdier vegetation forms which
adequately support nests through periods of severe weather (Francis 1971).
Buttonhush and bulrush were the most commonly used supportive plants
and nests placed in these plants also had relatively high success, 32% and 26%,
respectively, of all successful nests (Table 7). The grouping of all other

268

THE WILSON BULLETIN • Vol. 90, No. 2, June 1978

Supportive Plant Type

Table 7

AND Nesting Success

FOR 127 Active Nests

IN 1975

Supportive

Number

Number known

Percent

plant

of nests

successful

Buttonl)usli

Willow

Rush

Bulrush

Cattail

Common Alder

Others

Totals

127

seldom-used plants had a rather high success (47.1%) ; the reason for which
is not clearly understood. Robertson (1972 ) also observed a relatively high
success in rarely-used upland woody vegetation. He suggested that these
favorable nesting sites are not used more often because they are poorly suited
for the grouped territorial nesting pattern to which Red-wings are adapted.

Only 174 (22%) known unsuccessful nests were recorded out of 755 active
nests, from 1963 to 1975, compared to 82 ( 65% ) unsuccessful nests re-
corded out of 126 carefully observed active nests in 1975. In 1975, 88 in-
stances of predation occurred on the 82 nests. Nest losses not accompanied by
significant nest damage, an indication of probable snake or bird predation,
were most common ( 56 of 88 occurrences ) throughout the nesting season.
Mammalian predators played a lesser role (27); whereas weather (3) and
mowing or grazing ( 2 ) were rather insignificant contributors to lack of nest-
ing success.

Other losses can be attributed to the Brown-headed Cowbird {Molothrus
ater). Smith (1943) found the incidence of parasitism by the Brown-headed
Cowbird to be from 5% (1940) to 0.6% (1941) for the same location. We
found 12 of 755 (1.6%) active nests were parasitized from 1963 to 1975.
Sixteen cowbird eggs were laid, including 1 nest with 3 cowbird eggs, 2 nests
with 2 eggs, and 9 nests with 1 egg. In the 12 parasitized nests, an average
of 2.2 (range, 1-4) Red-wing and 1.3 cowbird eggs were present. Cowbird
eggs were found in Red-wing nests from 30 April to 29 May at heights of
1 to 2.5 m in a wide range of supportive plants in all the major habitat types.
The cowbird laying peak seemed to coincide with the peak laying period
of the Red-wing ( Goertz 1977 ) . In 1975, 2 cowbird eggs were measured,
being 15.8 X 20.8 mm and 16.4 X 19.8 mm, only slightly smaller than Red-
wing eggs.

Brown and Goertz • REU-WINGED BLACKBIRDS IN LOUISIANA 269

SUMMARY

Data were collected on 755 active nests of the Red-winged Blackbird {Agelaiiis
phoeniceus) in a largely forested area of north Louisiana from 1963 to 1975. Nesting
began in early April and continued until early August; the height of the nesting season
was during May. Completed clutches contained an average of 3.82 eggs (range, 2 to 5) ;
broods contained an average of 2.87 (range, 1 to 5j young. During 1975, 0.83 young
fledged per nest for 127 active nests.

Thirty species of plants were used as nest sites, with the ratios of plants used being
different in each of the 4 major habitats. Average nest height varied with the habitat
in which the nest was located. Open freshwater areas were the most heavily used
habitats. Nesting success was directly related to the height of the nest, depth of water
below the nest, type of vegetation used for nest support, and nesting habitat. Predation
was a major factor involved in nest loss.

ACKNOWLEDGMENTS

We are grateful to James G. Dickson and John L. Murad for assistance during certain
phases of this work, to Donald G. Rhodes for help in plant identification, to John E.
Carothers for help in making statistical comparisons, to many students who helped in
locating nests, and to Richard A. Dolbeer, Brooke Meanley, and Raleigh J. Robertson
for their review of this manuscript.

LITERATURE CITED

Allen, A. A. 1914. The Red-winged Blackbird: a study in the ecology of a cat-tail
marsh. Proc. Linnaean Soc. New^ York 24^25:43-128.

Beer, J. R., and D. Tibbitts. 1950. Nesting behavior of the Red-winged Blackbird.
Elicker 22:61-77.

Bent, A. C. 1958. Life histories of North American blackbirds, orioles, tanagers, and
allies. U.S. Natl. Mus. Bull. 211.

Brenner, E. J. 1966. The influence of drought on reproduction in a breeding popu-
lation of Red-winged Blackbirds. Am. Midi. Nat. 76:201-210.

Case, N. A., and 0. H. Hewitt. 1963. Nesting and productivity of the Red-winged
Blackbird in relation to habitat. Living Bird 2:7-20.

Darling, E. E. 1938. Bird flocks and the breeding cycle. Cambridge University Press,
England.

Dolbeer, R. A. 1976. Reproductive rate and temporal spacing of nesting of Red-
winged Blackbirds in upland habitat. Auk 93:343-355.

Fautin, R. W. 1941. Development of nestling Yellow-headed Blackbirds. Auk 58:
215-232.

Francis, W. J. 1971. An evaluation of reported reproductive success in Red-winged
Blackbirds. Wilson Bull. 83:178-185.

. 1973. Blackbird nest placement and nesting success. Wdlson Bull. 85:86-87.

. 1975. Clutch size and nesting success in Red-winged Blackbirds. Auk 92:

815-816.

Frankiiauser, D. P. 1964. Renesting and second nesting of individually marked Red-
winged Blackbirds. Bird-Banding 35:119-121.

Goddard, S. W., and V. V. Board. 1967. Reproductive success of Red-winged Black-
birds in north central Oklahoma. Wilson Bull. 79:283 289.

270

THE WILSON BULLETIN • VoL 90, No. 2, June 1978

Goehtz, J. W. 1977. Additional records of Brown-headed Cowbird nest parasitism in
Louisiana. Auk 94:386-389.

Gottsciialk, J. S. 1967. The federal role in dealing with the blackbird problem.
Pp. 26-32 in Proceedings of the North American conference on blackbird depreda-
tion in agriculture ( M. L. Giltz, Ed.). Neil House, Columbus, Ohio.

Holcomb, L. C., and G. Twiest. 1971. Growth and calculation of age for Red-winged
Blackl)ird nestlings. Bird-Banding 42:1-17.

Meanley, B. 1971. Blackbirds and the southern rice crop. Bur. Sport Fish. Wildl.
Resource Publ. 100.

Orians, G. H. 1961. The ecology of blackbird (Agelaius) social systems. Ecol.
Monogr. 31:285-312.

Robertson, R. J. 1972. Optimal niche space of the Red-winged Blackbird (Agelaius
phoeniceus). I. Nesting success in marsh and upland habitat. Can. J. Zool. 50:
247-263.

. 1973. Optimal niche space of the Red-winged Blackbird: spatial and temporal

patterns of nesting activity and success. Ecology 54:1085-1093.

Shaw, S. P., and C. G. Fredine. 1956. Wetlands of the United States. Fish and Wildl.
Serv. Circ. 39.

Smith, H. M. 1943. Size of lireeding populations in relation to egg-laying and re-
productive success in the Eastern Red-winged Blackbird {Agelaius phoeniceus).
Ecology 24:183-207.

St. Amant, L. S. 1959. Louisiana wildlife inventory and management plan. Louisiana
Wildl. Fish. Comm., Baton Rouge.

Stowers, J. F., I). T. Harke, and A. R. Stickley. 1968. Vegetation used for nesting
by the Red-winged Blackbird in Florida. Wilson Bull. 80:320-324.

Young, H. 1963. Age-specific mortality in the eggs and nestlings of blackbirds. Auk
80:145-155.

DEPT. OF ZOOLOGY, LOUISIANA TECH UiMV., RUSTON 71272. ACCEPTED 15
DECEMBER 1976.

Wilson Bull, 90(2), 1978, pp. 271-284

THE RUFOUS-COLLARED SPARROW AS A HOST
OF THE SHINY COWBIRD

Rosendo M. Fraga

The Rufous-collared Sparrow or Chingolo (Zonotrichia capensis) is one of
the main hosts of the Shiny Covvhircl [Molothrus boimriensis) through most
of their overlapping range in South America. The most recent study on the
interaction between these species was carried out by King (1973) in Horco
Molle, province of Tucuman, northwestern Argentina.

In 1970 I began to collect data on all the nests of Rufous-collared Sparrows
which I found in my study area near Lobos, province of Buenos Aires,
Argentina. Rufous-collared Sparrows are quite abundant here and I have
found 45 nests.

The data reported here relating to the interaction between host and brood
parasite differ from data collected from Tucuman by King (1973), and also
from the data collected near Rio de Janeiro, Brazil, by Sick and Ottow ( 1958;
original not seen: their observations are extensively quoted in the paper

of King) . My data are also relevant to other studies on the brood parasitism of
the Shiny Cowhird which 1 am carrying out in the same study area. One of
these studies, begun in 1972, is centered on the reproductive interactions of the
Shiny Cowhird and the Chalk-browed Mockingbird [Minius saturninus) .
I think that only comparative studies will throw light on some aspects of
the breeding biology of the Shiny Cowhird, such as the significance of the
polymorphism in egg shell color.

Since 1969 I have also been studying the breeding of the Bay-winged
Cowhird or Bay-wing {Molothrus badius) and its brood parasite the Scream-
ing Cowhird {Molothrus rujoaxillaris) . I have already published a short
preliminary account on both species (Fraga 1972). Since I began this study
I have found an increasing number of nests of Bay-wings parasitized by both
the Screaming and the Shiny cowhirds. This situation seems to have been
overlooked and was only briefly commented on by Friedmann (1929:119),
although among other things it is an additional complication to the problem
of the ])roper identification of the eggs found in the nests of Bay-wings.

STUDY AREA AND METHODS

All the nests of sparrows were studied in the main woodland of Estancia La Candelaria,
Lohos, Province of Buenos Aires. Argentina (about 35° 15' S) and in 2 smaller adjacent
wooded areas. Between 1970-1972 I found most nests in a small woodland of about 0.8
ha, fenc(‘d off from the surrounding pasture with barbed wire (the fenced area is about 1

272

THE WILJSON BULLETIN • l ol. 90, No. 2, June 1978

ha). The most abundant trees are the North American locusts Robinia pseudoacacia
and Gleditsia triacanthos, and the native tala {Celtis spinosa). As this area is protected
from grazing cattle and horses, young trees and saplings are quite numerous. Erom
September to early January there is a rich cover of tall grasses, chiefly Bromus unioloides.
Erom 1972 onwards most nests were found in the main woodland of La Candelaria, a
more complex area which covers about 80 ha. Basically this area includes a central zone of
gardens, orchards, and lawns surrounded by a belt of untended dense woodland where
several species of trees, shrubs, palms, and vines have become naturalized. In these
years I have also studied some nests in a hedgerow of native trees and shrubs ( Celtis
spinosa, Parkinsonia aculeata. and Sarnbucus australis) bordering a drainage ditch.

All nests were found along routes or in places which I visited frequently, often daily,
for the study of other birds. The sample of 45 nests could be divided into 2 subsamples.
The first comprises 40 nests which were found only with eggs of sparrows or cowddrds
or of both species. In this suhsample only 7 nests Avere found in or before the egg-
laying period of the sparrows, as new' sparrow eggs Avere laid after my first visit. The
remaining 33 nests were found during the incubation period of the sparrows. The second
subsample comprises 5 nests Avhich Avere found Avith at least one nestling sparroAv or
cowbird. The inclusion of this subsample in the computations of nesting success Avould
overstate success. My experience Avith both species suggests that non-parasitized nests
have a higher chance of reaching the nestling or fledgling stage than parasitized ones.
For this reason this subsample Avill be excluded from computations on the incidence of
parasitism.

Once found, all the nests of sparrows were visited either daily or at 2-day intervals.
All the eggs and the younger nestlings Avere marked; older nestlings were banded Avith
colored celluloid l)ands. Day 0 designates the day in which the nestlings hatched, and
day 1 is the folloAving day. Nestlings Avere usually Aveighed around noon.

THE LOCAL HOSTS OF THE SHINY COWBIRD

Broadly speaking we have good general information on the species of
passerines parasitized by Shiny Cow birds in this part of Argentina (Fried-
mann 1929, 1963 j though more intensive studies are needed. In 8 years of
bird study in La Candelaria I have found eggs of Shiny Cowbirds in the
nests of 12 species of passerines. About 33 species of passerines breed
regularly here, but my coverage for some species breeding exclusively in open
country or in marshes is not good. Most of my nesting records ( more than
350 up to 1975 ) are of the 24 species of passerines that breed in w ooded
areas. In this group I have found eggs of Shiny Cowbirds in the nests of the
following 8 species: Cattle Tyrant [Machetornis ri.xosus), Fork-tailed Fly-
catcher (Muscivora tyramms), Tropical Kingbird { Tyrannus melancholicus ) ,
House Wren [Troglodytes aedon). Chalk-browed ^lockingbird, Rufous-
bellied Thrush [Turdus rufiventris) , Bay-w inged Cowbird, and Rufous-
collared Sparrows. More than 50% of the nests of the Fork-tailed Flycatchers,
mockingbirds, and sparrows were parasitized. I paid particular attention to
the local species recorded as effective hosts (rearing fledgling cowbirds) by
Friedmann (1963:197), such as the Rufous Hornero ( Furnarius rufus) and

Fraga • RUFOUS-COLLAREl) SPARROWS

273

the Masked Gnatcatcher [Polioptila dumicola) ; for these species I have no
local records of parasitism.

Although this list could eventually he enlarged, the important point is that
it is unlikely that any of the remaining 16 species could be an important
effective host of the Shiny Cow bird. This is also show n by the following
list of 6 species of woodland birds that have been recorded rearing fledglings
of Shiny Cow birds in La Candelaria betw een 1970-1975: Cattle Tyrant (1
record), Fork-tailed Flycatcher (no less than 14 records), Tropical Kingbird
(2 records). House Wren (4 records). Chalk-browed Mockingbird (6
records), and Rufous-collared Sparrows (12 records excluding those re-
ported here ) . The lack of records for the Rufous-bellied Thrush may simply
reflect the fact that it is a scarce bird in my study area; the peculiar situation
of the Bay-wing needs another explanation.

BREEDING BIOLOGY OF RUFOUS-COLLARED SPARROWS

Rufous-collared sparrows are resident throughout the year in La Cande-
laria. Some males may show sporadic outbursts of territorial behavior (in-
cluding singing) in early August in periods of fair weather, l)ut I have not
detected nesting activities at such early dates. Sustained territorial behavior
can he observed from late August to early February.

My earliest record of a nest-building female is 16 September (1975) in
the main woodland. Table 1 shows the number of nests of sparrows by
periods of half months. The start of egg-laying was known or could be
calculated in 22 nests. The remaining 23 nests are placed in the table in the
period in which they were found. In my sample the earliest date for egg-
laying was 22 September 1975 (calculated). The latest nest (with eggs) was
found on 11 February 1971.

In La Candelaria all the nests of sparrows which I found were built on
or near the ground. The highest nest was built 12 cm al)ove ground level
in a tangle of the vine Doxantha unguis-cati growing over a casuarina tree.
Most nests were built in dense cover and near the bases of trees, fallen
branches, or protruding roots. Otherwise there was considerable variation in
the location and degree of exposure of the nest. In places with a sparse
ground cover, such as dense woodland, nests were often found under piles of
fallen branches. Some peculiar locations include 2 nests liuilt inside low, open
hollow stubs (both parasitized), another built in a cavity at the base of a tree,
2 nests built in deep niches in the slopes of a drainage ditch (1 parasitized),
and one nest found under the concave side of the basal sheath of a fallen
palm frond (parasitized). Some nests were comtiletely covered from above.

As most local passerines, Rufous-collared Sparrows lay either 3 or 4
eggs. The mean clutch size of 11 non-parasitized nests was 3.18 eggs; there

274

THE WILSON BULLETIN • VoL 90, No. 2, June 1978

Table 1

Temporal Uisthibution of Nests of Rufous-collared Sparrows,

IIALF-MONTIIS*

RY PERIODS OF

Period

Parasitized

nests

Non-parasitized

nests

Total

15-30 September

1-15 October

16-31 October

1-15 November

16-30 November

1-15 December

16-31 December

1-15 January

16-31 January

1-15 February

—

TOTAL

* Five nests found

in the nestling period are included

in the table.

were 9 nests with 3 and 2 nests with 4 eggs. Of the 3 non-parasitized nests
found in the nestling stage, 2 contained 3 nestlings and the other 4 nestlings.
A comparison with the data assembled by King (1973: Table 1) shows the
expected gradual increase in clutch size with latitude ( 2.69 eggs in Tucuman
and 2.31 eggs in Rio de Janeiro).

Measurements of 32 eggs of Rufous-collared Sparrows were as follows:
range 16.4-21.8 X 13.5-15.9 mm; mean and standard error: 19.28 ± 0.23
X 14.76 ± 0.09 mm.

The incubation period could only be determined in one nest with 3 eggs,
and was 13 days. In this nest the first 2 eggs hatched on day 12, but the
spread of hatching was less than 24 h.

Nestlings were fed insects by both parents. By day 2 the nestlings may
partially open their eyes when begging food; this becomes more noticeable
at day 3. By day 6 pinfeathers begin to emerge. The nestling period of
Rufous-collared Sparrows was 9-11 days (7 nestlings). At day 8 nestlings
may attempt to leave the nest.

Three banded fledglings remained in the parental territory for at least
19 days. They were fed at first by both parents, later probably chiefly by the
male. At this time (15 December 1974) the female was apparently attending
a second nest which I was unable to locate.

These observations suggest that Rufous-collared Sparrows may attempt
2 or even 3 broods in a season, but with their low nesting success probably
few pairs will succeed in raising more than one brood.

Fraga • RUFOUS-COLLAREU SPARROWS

275

THE BREEDING SEASON AND NEST-SEARCHING BEHAVIOR OF SHINY COWBIRDS

In La Candelaria, Shiny Cowbirds can be seen in variable numbers
throughout the year. As I have but a few banded individuals, I know little
about their local movements. Outside the breeding season they come to the
main woodland only for roosting. The largest flocks can be seen in winter.

In my sample of parasitized sparrow nests the earliest date for eggs of
Shiny Cowbirds was 26 September 1975. This is the earliest date I have
for cowhird eggs in the nests of any host. The latest parasitized nest of
sparrows was found on 27 January 1972; I have 2 unquestionable later
records of parasitism by Shiny Cowbirds in 2 nests of Bay-wings: 2 im-
maculate white eggs were laid on 5 and 7 February 1975. The overlap in the
breeding seasons of Rufous-collared Sparrows and Shiny Cowbirds was 87%
in my sample but this figure is probably a minimal estimate of the actual
overlap.

I have 14 records of female Shiny Cowbirds watching the movements of
Sparrows. Twice I had already found a nest of sparrows in the area; one
of these nests was parasitized the following day. The female cowbirds were
watching the movements of the sparrows from some convenient perch. At
times they uttered chattering notes, particularly when they flew away. On
11 October 1970 while I was watching a nest-building female sparrow, I
observed a female cowbird that remained no less than 17 min doing the same
from a nearby branch. Then she alighted near the nest but soon flew away
while the female sparrow was chipping. This nest was never finished. Twice
2 female cowbirds were seen searching in the same area at the same time.

Only twice did I observe male Shiny Cowbirds that followed nest-searching
females as they moved away. This does not necessarily mean that they were
helping the females in any way. In my opinion most records in the earlier
literature, particularly in Hudson (1920 1:75) of male Shiny Cowbirds visit-
ing nests of other birds with females in this part of Argentina were probably
caused by confusion with male Screaming Cowbirds. As I will show in my
account on the species, pairs of Screaming Cowbirds not only visit nests of
their hosts, the Bay-wings, but also nests of other birds. I have no records of
Screaming Cowbirds visiting nests of Rufous-collared Sparrows.

INCIDENCE OF PARASITISM

In my sample of 40 nests found either during the egg-laying or the in-
cubation period of the sparrows, 29 (72.5%) nests were parasitized. In the
additional sample of 5 nests found in the nestling period, 2 nests were
parasitized.

A seemingly significant difference between my observations and those re-

276

THE WILSON HLiLEETIN • Vol. 90, No. 2, June 1978

ported l)y Kinj? (1973) from Horco Molle can be observed in the temporal
distribution of non-parasitized nests. At Horco Molle both host and parasite
have slightly longer breeding seasons vvitb an overlap of 87%. Although the
overall incidence of parasitism was perhaps lower than in my study area ( 66%
vs 72.5% in the sample of 40 nests, or vs 69% in the sample of 45 nests but
difference not significant), at the peak of the breeding season of tbe sparrows
all their nests (100%) were parasitized. As can be seen in Table 1, in my
study area non-parasitized nests were more evenly distributed throughout
the breeding season of the sparrows and even at the peak of their breeding
season (period 15 October-15 January) only 27 out of 35 nests (77.1%)
were parasitized.

SEQUENCE OF EVENTS IN FIVE PARASITIZED NESTS

Of the 7 nests found in or before tbe egg-laying period, 5 nests were para-
sitized in the following way:

Nest 11: 27 October 1972, found with sparrow egg; 28 October 1972,
this sparrow egg was removed, now one sparrow egg plus 2 cowbird eggs ( 1
immaculate, 1 spotted ) ; 29 October, now 2 sparrow eggs plus 4 cowbird eggs
(2 immaculate, 2 spotted) ; 30 October, 2 sparrow eggs (number 3 punctured)
plus 5 cowbird eggs (3 immaculate, 2 spotted ) but nest abandoned, all the
eggs were cold. The nest retained all its contents till 2 November, when 1
collected all the eggs. The nest was apparently parasitized by at least 2
female cowbirds.

Nest 15: 1 October 1973, found with 2 sparrow eggs; 2 October, 3

sparrow eggs; 3 October, 3 sparrow eggs plus 1 immaculate cowbird egg;
5 October, 1 sparrow egg apparently punctured, otherwise no changes; 6,
8, 10 October, no changes; 12 October, all eggs vanished.

Nest 27: 24 September 1975, 1 sparrow egg; 26 September, 2 sparrow
eggs plus 1 immaculate cowbird egg stained with yolk; 28 September, all eggs
gone.

Nest 34: (the highest nest I found in the creeper Doxantha) 11 November
1975, 2 sparrow eggs; 12 November, 3 sparrow eggs; 14 November, no
changes; 16 November, 3 sparrow eggs plus 1 spotted cowbird egg; 24^25
November, all sparrow eggs hatched; 28 November, cowbird egg hatched but
nestling died.

Nest 36: Found on 23 November 1975 without eggs; 24 November, no
eggs; 25 November, 1 immaculate cowbird egg; 26 November, 1 sparrow
egg plus the cowbird egg: 27 November, 2 sparrow eggs plus the cowbird
egg; 28 November, 1 sparrow egg (number 3, as both 1 and 2 removed)
plus 2 cowbird eggs (1 immaculate, 1 spotted); 29 November, 1 sparrow
egg (the same) plus 3 cowbird eggs (1 immaculate, 2 spotted) but nest

Fraga • RUFOliS-COLLARED SPARROWS

277

deserted. This nest retained all its contents till 3 December, then only the
cowhird eggs till at least 18 December. The nest was probably parasitized
by 2 female cowbirds.

The time of laying of the 11 cowhird eggs was as follows: 1 egg ap-

parently laid before the egg-laying of the sparrows; 7 eggs laid during the
egg-laying period; 2 eggs laid on the first day of the incubation period, and
1 egg laid on the fourth day of the incubation period. Sparrow eggs were re-
moved either on the day when a cowhird egg was laid or on the previous or
following day.

CONTENTS OF PARASITIZED NESTS

The 29 parasitized nests contained 59 cowhird eggs ( average 2.03 eggs
per nest) and 71 sparrow eggs which were eventually reduced to 62 (average
2.14 eggs per nestj. The average number of cowhird eggs per parasitized nest
in my sample was similar to the average of 2.06 cowhird eggs reported by
King (1973: Table Ij from Tucuman.

As the mean clutch size in non-parasitized nests of sparrows was 3.18
eggs, this means that cowbirds removed an average of 1.04 sparrow eggs per
parasitized nest, apparently a higher number than at Horco Molle or Rio de
Janeiro (King, 1973: Table Ij. 1 estimate that 30 eggs of sparrows were
removed by cowbirds in my sample. 1 actually observed the disappearance of
9 sparrow eggs from 7 nests. In addition no fewer than 6 sparrow eggs were
punctured hut not removed.

1 have no definitive records of cowhird eggs removed either by cowbirds
or by sparrows. However 1 have observed 2 punctured cowhird eggs in 2
nests with 4 cowhird eggs each ; both nests were abandoned on my next visit.

In my sample 15 parasitized nests (51.7%) contained 1 cowhird egg and
2-4 sparrow eggs (average 2.73 eggs); 4 nests (13.8%) contained 2 cow-
hird eggs and 2-3 sparrow eggs (average 2.75 eggs j ; 6 nests (20.7%) con-
tained 3 cowhird eggs and 0-1 sparrow eggs (average 0.66 eggs); 2 nests
(6.9%) contained 4 cowhird eggs and 2 sparrow eggs, and 2 nests (6.9%)
contained 5 cowhird eggs and 0-2 sparrow eggs (average 1 egg). Thus the
highest compound clutch was 7 eggs.

Cowhird eggs. — In my study area the eggs of Shiny Cowbirds are either
immaculate or spotted (Fig. 1); this is normal in eastern Argentina,
Uruguay, and southeastern Brazil (Friedmann 1929). Most immaculate eggs
are white; the ground color of the spotted eggs as well as the number, size,
distribution, and color of the spots shows a considerable and probably con-
tinuous variation. Intermediates to these basic types rarely occur. A few
immaculate eggs may have some extremely faded pale gray spots resembling
water marks on a white paper. These eggs were considered immaculate. In

278

THE WILSON BULLETIN • Vol 90, No. 2, June 1978

Fig. 1. Parasitized nests of sparrows. A: 2 sparrow eggs, 1 immaculate cowbird egg.
B: 2 spotted and 1 immaculate cowbird eggs. C: 1 sparrow egg and 3 rather similar
spotted cowbird eggs. D: 3 immaculate cowbird eggs; later 2 spotted cowbird eggs were
also laid in this nest.

the nests of Rufous-collared Sparrows, only 1 cowbird egg was considered
to be intermediate, having a white ground color with 7 small deep brown
spots. In my experience immaculate and spotted eggs are better regarded as
discontinuous or quasi-discontinuous forms. This basic variation seems to
be a true genetic polymorphism (Ford 1965). The possible selective forces
that maintain this polymorphism are unknown. Immaculate eggs are strik-
ingly different from the eggs of the local hosts of the Shiny Cowbird.
Hudson (1920 1:124-126) suggested that some host species may selectively
reject or eject the immaculate eggs. At least 3 species among the recorded
local hosts may eject cowbird eggs and at the present time I am investigating
this point in the Chalk-browed Mockingbird. The Rufous-collared Sparrow
is a poor subject for such studies, as it accepts eggs of both types.

Of the 59 eggs of Shiny Cowbirds found in nests of sparrows, 32 (54%)
were immaculate, 26 (44%) were spotted, and 1 was intermediate. The

Fraga • RUFOUS-COLLAREl) SPARROWS

279

Fig. 2. Roundness (width X 100/length) of immaculate and spotted Shiny Cowhird
eggs.

distribution of immaculate and spotted cowhird eggs in the parasitized nests
was as follows: 10 nests received only immaculate cowhird eggs (7 with 1
egg, 3 with 2 eggs I : 8 nests received only spotted cowhird eggs ( 7 with 1
egg, 1 with 3 eggs), and 10 nests were parasitized with both egg types in
various combinations. One nest was parasitized with 1 intermediate egg.
In the whole sample, 21 nests received immaculate eggs and 18 nests received
spotted eggs. Available evidence suggests random placement of both egg
types.

I measured 41 cowhird eggs (22 immaculate, 18 spotted and 1 inter-
mediate). Average measurements (mean and standard error) are 22.75
0.22 X 18.16 ± 0.18 mm for all the sample. Measurements of immaculate
eggs are as follows: range: 20.7-24.3 X 16.8-19.8 mm: mean and standard
error 22.56 ± 0.20 X 18.35 0.19 mm. Measurements of spotted eggs are:

range: 20.7-24.7 X 16.6-18.9 mm; mean and standard error: 22.91 ± 0.25 X
17.91 ± 0.17 mm. The intermediate egg measured 23.9 X 18.3 mm. As can he
observed immaculate eggs are usually more rounded. The index for round-
ness (width X 100/length) ranges from 76.82 to 84.75 for immaculate eggs
and 75.30 to 80.53 for spotted eggs. These values are plotted in Figure 2.
The difference in roundness between the immaculate and the spotted eggs
found in nests of sparrows is significant ( P < 0.001, Mann-\\ hitney L-test
and Kolmogorov-Smirnov test).

On the whole spotted eggs of Shiny Cowhirds do not closely resemlile those
of their hosts. However the spotted eggs are so varialile that I have found a
few exceptions in the nests of 4 host species. The smallest spotted cowhird egg

280

THE WILSOxN BULLETIN • Vol. 90, No. 2, June 1978

in my sample resembled the eggs of sparrows not only in size (20.7 X 16.6
mm I but also in color and pattern.

Incubation period. — In 2 parasitized nests of sparrows, 2 eggs of Shiny
Cowbirds hatched in 12 days. Six eggs in nests of other hosts hatched in
llVij (1 egg) and 12 days. Four fresh eggs of cowbirds (2 from deserted
nests of sparrows) placed in nests of Eared Doves {Zenaida auriculata)
hatched in 12 days.

DESERTION OF PAR.\SITIZED NESTS

I did not observe Rufous-collared Sparrows to remove cowbird eggs of any
type nor to practice egg-burial, but they may desert parasitized nests. The
only possible case of desertion of a non-parasitized nest was attributed to
predation. One sparrow egg was gone on the day of the desertion; the follow-
ing day the destroyed nest cup was empty. I agree with Rothstein ( 1975)
that nest desertion is not necessarily a direct specific response to cowbird eggs.
The parasitized nests were perhaps deserted because the female cowbird was
discovered by the sparrows, or because of the abnormal size of the clutch,
or because some punctured eggs became glued to the nest lining.

As nests of Rufous-collared Sparrows suffer a high rate of predation, it is
convenient to have a restricted definition of desertion. In this study I con-
sidered that 7 out of 29 parasitized nests were deserted due to cowbirds
because: (a) eggs of host and parasite, or of both species, remained in the
deserted nest, (b) the remaining eggs were not warmed and no sparrow was
seen in or near the nest, ( c) the nests retained their final contents for at least
1 day after their abandonment, or ( d) I have evidence that cowbirds visited
these nests either on the day of their desertion or ( in 1 instance) on the
previous day because new cowbird eggs were found or sparrow eggs were
removed or punctured. The 7 deserted nests contained 20 cowbird eggs (11
immaculate, 9 spotted; average 2.85 eggs) and 13 sparrow eggs (average
1.85 eggs) and included the 2 nests with 4 cowbird eggs and one nest with
5 cowbird eggs.

INTERACTIONS IN THE NESTLING PERIOD

Due to the low nesting success only a limited number of eggs produced
fledglings. In this section data from the additional 5 nests found in the
nestling period are also included.

Cowbird nestlings. — Nestling Shiny Cowbirds from the nests of 6 host
species showed variation in the color of the oral flanges and mouth lining,
and in fewer cases in the color of the skin. The color of the flanges varied
from pure white to yellow. The color of the mouth lining varied from pale

Fraga • RUFOUS-COLLARED SPARROWS

281

pink to deep red or orange-red. In at least 3 nestlings (1 from a nest of
sparrows ) the hue of the skin was yellowish.

These seemingly continuous variations were not obviously correlated with
egg shell color. I have not detected correlation between the colors of the
flanges and mouth lining. The fledgling cowhirds which were successfully
reared by sparrows could he better described by comparison with the nestling
sparrows, which have pale yellow flanges and red mouths. In 2 fledglings
the flanges were white; I had a pale red mouth. The other fledglings had
flanges which were at least as yellow as those of the sparrows; their mouth
linings were red. The colors of the mouthparts were apparently irrelevant to
the parental behavior of the sparrows. My largest sample of nestling Shiny
Cowhirds comes from parasitized nests of mockingbirds, in which they often
die of starvation; despite this, I have no conclusive evidence relating mouth
color to survival value.

By day 3 cowbird nestlings have partially open eyes; by day 7-8 pin-
feathers begin to emerge. Nestling cowhirds are less precocial in their
behavior than nestling sparrows of the same age and do not attempt to leave
the nest until day II or 12. The recorded nestling period of Shiny Cowhirds
in nests of sparrows was 12-13 days hut probably not all the nestlings de-
parted spontaneously. In the safer nests of other hosts, the recorded nestling
period was 13-15 days. Nestling Shiny Cowhirds seem to be less well
adapted to the hazards of ground nesting than are nestling sparrows.

Table 2 gives the weights of nestling cowhirds and sparrows. King ( 1973)
observed that the weight of a nestling cowbird is approximately equivalent
to the weight of 2 nestling sparrows of the same age. In my study area I
observed a similar situation. I found a pair of sparrows that fledged 4 young
sparrows (in a nest found with nestlings), thus one might expect that 2 cow-
hirds could also be fledged at times. However in 3 nests in which more than

1 cowbird egg hatched only I nestling cowbird survived. Events in these
nests were as follows:

Nest 21: found 27 October 1974 with 3 cowbird eggs (2 immaculate,

2 spotted ) and I sparrow egg; 30 October, the spotted cowbird egg hatched;
31 October, the 2 immaculate cowbird eggs hatched ( weights of nestlings —
5.4, 3.4, 3.0 gj ; I November ( weights 6.9, 4.4, and 3.7 g) ; 2 November (13,
6.8 and 5.8 g) ; 3 November, the 2 smaller nestlings were gone (the survivor
weighed 18 g ) ; 4 November, the nest was destroyed, the nestling vanished.

Nest 23: found on 8 December 1974 with 3 sparrow eggs and 2 cowbird
eggs (I immaculate, I spotted); 18 December, 2 cowbird nestlings (4.0, 3.7
g) ; 19 December, 2 cowbird nestlings (6.8 and 6.4 g) ; 20 December, 2
cowbird nestlings (8 and 7.7 g ) plus I sparrow nestling ( 1.8 g ) ; 21 Decem-
ber, only I cowbird nestling was alive (10.5 g) ; the other was dead from

282

THE WILSON BULLETIN • VoL 90, No. 2, June 1978

Table 2

Weights H;) of Nestling Sparrows and Cowrirds*

Day

Sparrows

alone

Sparrows with
cowbirds

Cowbirds

2.3**

2.0

3.7

( 1.7-2.6, N 9)

(1.9-2.1, N = 4)

(3-4.4, N = 5)

3.4

3.5

6.1

(3.2-3.B, N = 6)

(3.2-3.7, N =: 5)

(5.2-7.6, N = 5)

5.3

5.1

9.8

(5.1-5.7, N = 6)

(5-5.3, N = 4)

(7.6-10.5, N = 6)

7.7

7.6

12.4

( 7.2-8, N = 9 )

(7.4-7.8, N z= 4)

(11-15.5, N = 5)

10.1

9.6

18.8

(9.2-10.5, N = 9)

(9-10.5, N z= 4)

(14.5-23, N = 5)

12.2

23.0

(12.5-13.5, N = 6)

(12-13, N = 4)

(20..5-28, N = 5)

13.6

14.5

26.2

(13-14.5, N = 6)

(14-15, N = 4)

(23.5-29, N = 4)

16.7

31.2

(15.5-18, N = 6 )

(15.5-17, N r=4)

(27.5-34, N = 4)

16.4

31.8

(15.5-18.5, N = 3)

(28-34.5, N = 4)
32.2

(29.5-35, N = 2)

* Nestlings that died of starvation are excluded.

** Mean with range and sample size in parentheses.

unknown causes. The nestling sparrow was gone. Only 1 cowbird was
fledged.

Nest 28 B; found on 12 October 1975 with 2 nestling cowbirds (weights:
12.5, 6 gj ; 15 October (weights 20 and 13.5 g) ; 17 October (27 and 16 g) ;
18 October, only 1 nestling (29.5 g) which was fledged on 22 October.

In 2 parasitized nests, 2 nestling sparrows were raised with 1 nestling
cowbird. One nestling sparrow died of starvation in each of these nests.
Another fledged cowbird was raised alone (no sparrow egg hatched).

NESTING SUCCESS

Sparroiv eggs in n on-parasitized nests. — Of the 35 eggs, 26 were taken by
predators. Only 9 eggs hatched in 3 nests. Of the 9 nestlings, 3 were taken
by predators; 6 were fledged. Only 2 of 11 nests were successful (18.2%),
9 of 35 eggs hatched (25.7%), 6 of 9 nestlings fledged (66.7%), and 6 of
35 eggs produced young which fledged (17.1%).

Fraga • RUFOUS-COLLARED SPARROWS

283

Sparrow eggs in parasitized nests. — I estimate that 92 eggs were laid in
29 nests; 30 eggs were removed by cowbirds. Another 13 eggs were aban-
doned in deserted nests, 23 eggs were taken by predators. Only 13 eggs
hatched in 5 nests. Of the 13 nestlings 3 died of starvation, 4 were taken by
predators, and only 7 were fledged. Only 3 of 29 nests were successful
(10.3%), 13 of 92 eggs hatched (14.1%), 7 of 13 nestlings fledged (53.8%),
and 7 of 92 eggs produced young which fledged (7.6%).

Sparrow eggs in all nests. — I estimate that only 13 sparrows fledged from
127 eggs laid (10.2%) in 40 nests.

Cowbird eggs. — Of the 59 eggs, 20 were abandoned in deserted nests.
Another 27 eggs were taken by predators. Only 10 eggs hatched. Of the 10
nestlings, 3 died of starvation, 1 died from an unknown cause, and 2 were
taken by predators. Only 4 cowbirds were fledged. Four of 29 nests were
successful (13.8%). Of 59 cowbird eggs laid, only 10 hatched (16.7%), and
of these nestlings, only 4 fledged (40%).

The figures for nest success are considerably lower than those reported
from Tucuman and Rio de Janeiro (King 1973, Table 1). They are also the
lowest figures for nesting success I have observed in local birds, but 1 have
not studied other ground nesters. As most nests of sparrows in my sample
were not found before egg-laying, the true nesting success could be even lower.
My figures suggest that Shiny Cowbirds probably surpass any nest predator
in the amount of harm inflicted to the reproductive efforts of the sparrows.

In my study areas nests built above the ground in trees and shrubs are
exposed chiefly to avian predators and to the only abundant climbing mam-
mal, the white-eared opossum i Didelphis albiventris) . Nests built on or near
the ground are probably equally exposed to opossums and to the most
abundant avian predators (such as the Chimango, Milvago chimango) but
in addition they are also exposed to other terrestrial predators, ranging from
amphibians (the escuerzo, Ceratophrys ornata) to foxes { Dusicyon gymnocer-
cus ) . Some terrestrial predators frequently seen in my study areas were
hog-nosed skunks [Conepatus chinga) and great tegu lizards iTupinambis
teguixin) .

The value of the sparrows as hosts of the cowbirds is considerably dimin-
ished by their high nesting losses. More extensive comparisons with other
host species will be included in future accounts.

SUMMARY

Brood parasitism of Shiny Cowbirds on Rufous-collared Sparrows was studied in
1970-1975 in north Buenos Aires province, Argentina. Summary information on the
local hosts of Shiny Cowd)irds is reported. The overlap in the breeding seasons of
cowbirds and sparrows was at least 87%. Shiny Cowlnrds parasitized 72.5% of the
nests. Non-parasitized nests were found almost throughout the breeding season of the

THE WILSON BULLETIN • Vol. 90, No. 2, June 1978

2VA-

sparrows. 01)servations on tlie nest-searching behavior of female cowhirds are included.
The sequence of egg-laying in 5 parasitized nests is described.

Parasitized nests contained an average of 2.03 cowhird eggs and 2.14 sparrow eggs.
Cowhirds removed al)out 1 sparrow egg per parasitized nest. More than 50% of the
nests were parasitized with 1 cowhird egg. Up to 5 cowhird eggs were found in para-
sitized nests. Cowl)ird eggs in nests of sparrows were immaculate ( about 55% ) or
spotted (about 45%). Immaculate eggs are more rounded. Rufous-collared Sparrows
deserted some parasitized nests.

The sparrows have a low nesting success i)robably because they are ground nesters.
Eew nests produced fledglings. Data on weights of nestlings suggest that sparrows
cannot rear more than 4 nestling sparrows or 2 nestling cowhirds, but the latter
situation was not observed.

SUMARIO

El parasitismo de erfa de Molothrus bonariensis sobre Zonotrichia capensis fue
estudiado durante 1970-1975 en una zona de estudio en Buenos Aires, Argentina. La
superposicion de perfodos de postura fue de 87% y el porcentaje de nidos parasitados
72.5%. El promedio de huevos hallados en dichos nidos fue 2.03 para el parasite y 2.14
para el huesped. Los huevos del parasite eran inmaculados o manchados, con diferencias
en la forma. Pese al bajo exito reproductive los dates sugieren que Zonotrichia puede
criar hasta 4 pichones propios o 2 del parasite, aunque lo ultimo no fue observado.

ACKNOWLEDGMENTS

I am most grateful to the editor and the reviewers for helpful comments and im-
provements in the manuscript and to Bette J. Schardien for preparing Figure 2.

LITERATURE CITED

Ford, E. B. 1965. Genetic polymorphism. M.I.T. Press. Cambridge, Massachusetts.
Fhaga, R. M. 1972. Cooperative breeding and a case of successive polyandry in the
Bay-winged Cowhird. Auk 89:447-449.

Friedmann, H. 1929. The cowhirds. C. C. Thomas. Springfield, Illinois.

. 1963. Host relations of the parasitic cowhirds. U. S. Natl. Mus. Bull. 233.

Hudson, W. H. 1920. Birds of La Plata. Vol. 1. J. M. Dent and Sons Ltd., London.
King, J. R. 1973. Reproductive relationships of the Rufous-collared Sparrow and the
Shiny Cowhird. Auk 90:19-34.

Rotiistein, S. I. 1975. An experimental and teleonomic investigation of avian brood
parasitism. Condor 77:250-271.

CALLAO 1502-4°, 1024 BUENOS AIRES, ARGENTINA (PRESENT ADDRESS: GUIDO
160o, 16 B-1016 BUENOS AIRES, ARGENTINA). ACCEPTED 12 NOV. 1976.

Wilson Bull, 90(2), 1978, pp. 285-287

GENERAL NOTES

Sexual similarity of Red-headed Yi oodpeckers and possible explanations based
on fall territorial behavior. — As discussed by Goodwin (Bull. Br. Mus. Zool. 17:1-44,
1968) the sexes are alike or nearly so in only 5 species of woodpeckers. In none of these
are the sexes more exactly alike than in Red-headed Woodpeckers ( Melanerpes erythro-
cephalus). This presents a challenging problem that has received little attention.

As narrated elsewhere, (Kilham, Wilson Bull. 70:347-358, 1959) 12 Red-headed Wood-
peckers settled in one small wood of 1.7 ha in Maryland attracted by pin oak ( Quercus
pcihistris) acorns. The wood was divided into 12 sharply defined territories, each wood-
pecker defending its stores, chiefly against interspecific intruders.

In such situations, I suggest that the monomoiphism of Red-headed Woodpeckers aids
females in establishing and maintaining individual winter territories. If males dominated,
the females would be crowded into less favorable habitats. This in turn might mean a
poorer winter survival. If, however, as may have happened in their evolutionary past,
females were selected to resemble males in plumage and hence have the same display
colors, they would have a more equal chance in border contests.

The best parallel that I have been able to find for the sexual similarities in color
of M. erythrocephalus is that described by Lack (Life of the Robin, H. F. and G. Witherby
Ltd., London, 1943) for the British Robin (Erithucus rubecola) . These birds form
small, individual fall territories and the sexes have identical coloration.

One might ask how do juveniles before molting to adult plumage fare in competition
with adults? As noted elsewhere (Kilham, op. cit.), among the 12 closely adjacent
winter territories observed, the 3 held by juveniles were all peripheral and appeared to
be the least desirable. The juveniles, therefore, without red heads, appeared to have
fared less well, but lack of experience may also have been a factor.

It would appear from descriptions by Bock (Univ. of Calif. Publ. Zool. 92:1-100,
1970) that the Lewis Woodpecker (Asyndesmus lewis) resembles M. erythrocephalus in
being irregularly migratory in relation to fall storage territories. This may account for
the similarity in plumage between the sexes of this species. Acorn Woodpeckers (.17.
Jormicivorus) , which are sexually dichromic, also store mast in the fall. Living in
social groups, however, and being to a considerable extent resident on the same territories
the year around (MaeRoberts, Condor 72:196-204, 1970) they are not exposed to the
same selection pressures as are A. lewis and M. erythrocephalus.

Among sapsuckers, as well presented by Howell (Condor 54:237-282, 1952; Auk
70:118-126, 1953), the eastern Yellow-bellied ( Sphyrupicus v. varius) is dichromic and
highly migratory whereas the western race, S. v. ruber, is monochromic and essentially
non-migratory. Are there any parallels to the situation encountered in .17. erythro-
cephalus? In the absence of information as to whether S. v. ruber maintains fall and
winter territories, I find it difficult to draw conclusions. It seems likely that mono-
chromatism in birds can arise from more than one kind of selection pressure and that
which I have described for an acorn-storing species of woodpecker may not apply,
necessarily, to other Picines with other habits. — Lawhence Kilham, Dept, of ,17/cro-
biology, Dartmouth Medical School, Hanover, NH 03755. Accepted 5 Feb. 1977.

Wilson Bull, 90(2), 1978, pp. 28.5-287

Notes on the courtship h<*havior of Brown-capped Rosy Finches. — Published
observations on the life history of the Brown-capped Rosy Finch i Leucosticte australis)

285

286

THE WILSON BULLETIN • VoL 90, No. 2, June 1978

are scanty, the most comprehensive Ijeing those of Bailey and Niedrach (Birds of
Colorado, Denver Mus. Nat. Hist., Denver, 1965), and Packard iin Bent, U.S. Natl.
iMus. Bull. 237, part 1, 1%8). These authors point to the lack of data on early breeding
season activities and the courtship behavior of this species. In order to help fill the hiatus
in our knowledge of the biology of these interesting birds, I offer the following notes on
their courtship behavior.

Courtship display.- — On 11 July 1977 1 was looking for rosy finches in the upper portion
of Navajo Basin (3700 m elev.j in the San Miguel Range, Dolores County, southwestern
Colorado. At 10:45 a female finch landed in the fellfield near me and began to forage.
Less than a minute later a male finch landed on a rock near the female and began to
court her. His body was held low, feathers fluffed, tail held high above the back, head
and neck stretched slightly forward and level with the back, throat swollen as he
emitted a continuous chirping and rapidly fluttered his wings at his sides. He displayed
for about 2 min then hopped off the rock and walked across the tundra toward the
female. He continued to display as he followed her. His wing-fluttering became more
pronounced, and he held his wings partly unfolded and away from his sides so that
they caught the wind, making him appear off balance. The female then moved closer
to the male, whereupon he flew at her and chased her in a horizontal straight line 1 m
above the ground for 100 m before they disappeared behind some boulders, terminating
the observations.

On 21 July 1977, I encountered many rosy finches around the base (3500 m elev.) of
a prominent buttress above Stillwater Reservoir on the northeast edge of the White
River Flattops in the Flattops Wilderness Area, Garfield County, northwestern Colorado.
At 10:10 a male began to display before a female in the basalt talus about 10 m below
me. The appearance of the display was like that described above with one major addi-
tion; the male picked up several long pieces of dried grass in his beak and held the
material as he chirped and fluttered his wings like a begging juvenile. This display
lasted about 2 min before the female showed her disinterest in the male’s efforts by
flying off.

On 22 July 1977, I visited Medicine Bow Peak in the Snowy Range, Albany County,
southeastern Wyoming, and found Brown-capped Rosy Finches along the base of the
south face (3300 m elev.). At 16:45 a male followed a female into the boulders and began
to display while perched atop a 2 m diameter rock as the female foraged below him.
Again, his physical appearance and actions were similar to those described above. He
picked up a piece of dried grass as he moved among rocks in rapid pursuit of the female.
He dropped the grass when he reached another boulder, ceasing to display about 5 sec
later when the female moved rapidly to the far side of the boulder out of his vision.
At 17:00 I saw another male follow a female into the boulders and begin to display,
but he barely had opportunity to begin when she apparently eluded him. He ceased
to display immediately and sat silently for several minutes atop the boulder with his
feathers fluffed.

The courtship display of the Brown-capped Rosy Finch seems to be similar to that
of the Black Rosy Finch (L. atrata) . The description provided by French (Auk
76:159-180, 1959) varies from what I saw' in the Brown-capped males only in the
elevated position of the beak. I saw nothing suggesting a female display. No coition was
seen, and the late dates of these observations coupled with the lack of response shown
by the females (except in the first case) may indicate that these males were courting
already-mated females making foraging trips for their young (I saw fledged juveniles
being attended by their parents at the last 2 localities). The persistent nature of un-

GENERAL NOTES

287

mated male rosy finches is well documented (French, op. cit.; Johnson, Auk 82:190-205,
1965; Twining, Condor 40:246-247, 1938).

An analysis of the courtship display described above seems premature, therefore I
wish only to suggest a possible origin of the display based on my first impression. Hinde
(Ibis 97:706-745, 1955; 98:1-23, 1956) summarizes an analysis he made of the court-
ship behavior of several species of finches, in which he concludes the male courtship
displays are modified forms of the head-forward threat posture. My impression of the dis-
play of the male rosy finches is its similarity to the juvenile begging response, particu-
larly the fluffed feathers, wing fluttering, and constant chirping. Morris (Behaviour
9:75-113, 156) mentions that in Estrildine finches feather postures are used as
social signals, and a fluffed body posture can eliminate normal individual distances
maintained by conspecifics, thus allowing individuals to approach one another until
touching, without fear of attack. It seems possible that the display used by a male rosy
finch acts to neutralize the expected agonistic response of the female upon his approach,
and may even invite her closer approach. Male aggressiveness during courtship at-
tempts, which Hinde (op. cit.) documents in a variety of Fringillids, may be more
readily apparent earlier in the season.

Flight display. — Packard (op. cit.) includes observations by R. J. Niedrach of a “con-
spicuous song flight” that occurs during the mating season. I have observed this “song
flight” perhaps a dozen times. In undulating fashion the male Brown-capped Rosy Finch
flies a large horizontal arc or circle traveling several hundred meters, chirping (described
in my field notes as a guttural churk) as he flaps his wings. During the breeding season
rosy finches are often scattered throughout a cirque or basin making it difficult to
visually locate conspecifics for courtship. The “song flight” of rosy finches may have
the same function as the advertising song of other species (see Tinbergen, Trans. Linn.
Soc. N.Y. 5:1-94, 1939) in that it serves to attract or locate potential mates. Finches
on the ground often respond vocally to others flying overhead or nearby.

The vocal flight display of Brown-capped Rosy Finches may have evolved due to other
selective parameters of the alpine environment. Morton (Am. Nat. 108:17-34, 1975)
presents an argument for the ecological selection of non-ground song displays used by
many grassland and tundra birds based on the effects of wind turbulence and solar
radiation on sound propagation from the ground. It seems possible that the acoustic
properties of an alpine environment, subjected to similar wind and temperature effects as
grasslands and tundra, may have a similar selectivity for aerial song displays in alpine
nesting birds. It is interesting to note that Horned Larks ( Eremophilci olpestris) and
Water Pipits (Anthus spinal ett a) , both of which nest in alpine areas where rosy finches
are found, also have flight displays (Verbeek, Wilson Bull. 79:208-218, 1967; Verbeek,
Auk 87:425-451, 1970).

My field studies wxre funded in part by a Margaret Morse Nice Award generously
provided by the Wilson Ornithological Society. I thank Drs. John P. Hubbard and
Phillip L. Wright, and the reviewers of this paper for their help and suggestions, and I
especially thank my parents for their continued support. — Paul Hendricks, 305 East
Maplewood Ave., Littleton, CO 80121. Accepted 21 Feb. 1978.

Wilson Bull, 90(2), 1978, pp. 287-290

Effects of nest removal on Starling populations. — Starlings iSturnus vulgaris)
commonly compete with W'ood Ducks i Aix sponsa) for nesting boxes. Bellrose and
McGilvrey (Wood Duck Management and Research: A Symposium, pp. 125-131. Wildl.

288

THE WILSON BULLETIN • Vol. 90, No. 2, June 1978

Manage. Inst., Wash., D.C., 1965) reported Starlings destroyed 23.8 and 20.6% of Wood
Duck nests in Illinois boxes in 1963 and 1964 respectively, and usurped a large number
of other boxes before Wood Ducks could use them. A similar situation has been re-
ported by Muncy and Burbank in Tennessee ( Proc. Southeastern Assoc. Game and Fish
Commissioners 29:493-500, 1975).

We conducted a Wood Duck nesting study at 10 sites in eastern Massachusetts during
1967-1973. Boxes were checked every 7 to 10 days between 1 April and 4 July and
Starling nests removed. Starlings are persistent nesters. If a nest with a partially com-
pleted clutch is destroyed, the hen is capable of reconstructing the nest within a day
and will frequently complete and incubate her remaining clutch (Kessel, Am. Midi.
Nat. 58:257-331, 1957). If an incubated clutch is destroyed, the hen will begin a new
clutch in 6-9 days ( Royall, Condor 68:196-205 ), 1966). Consequently, we removed many
more nests than there were pairs of Starlings present. Since individual Starlings were
not marked, we did not know how many pairs may have used a given box. Observations on
the stage of nest and clutch completion from 1 week to the next indicated that more than
1 pair of Starlings used some boxes. Once incubating Starlings were captured in a box
and bird and clutch destroyed each week for 3 consecutive weeks. There was a 4th
clutch the following week.

Figure 1 graphs the number of duck boxes used by Starlings and the total number
of nests removed during the 1967-1973 period and for 3 years following when boxes were
checked irregularly. It also shows the number of boxes used by Wood Ducks and
Hooded Mergansers ( Lophodytes ciicuUatus ) during the same time span.

Starlings used 47 boxes on 7 sites in 1967 and 146 nests were removed. Nesting
stages varied from nearly completed nests to completed clutches being incubated. Eight
sites had Starling nests in 1968 and all 10 sites were used in 1969. The first results of
the Starling nest removal program were noted in 1970 on a pond in Holden, Massachu-
setts. Starlings used 3 boxes on the pond in 1967, 2 in 1968, 1 in 1969 and did not nest
in any box thereafter. Starlings nested on 6 sites in 1971, 4 in 1972, and in 1973 only 3
sites had Starlings; 21 nests were removed from 10 boxes in 1973.

During 1974^1976, Starling nests were removed when encountered, but checks were
infrequent. As a result, a few broods were successfully fledged. Nests were started
on 6 areas in 1974 and on 7 in 1975, hut only 4 of these areas were used both years.
The same 4 areas were the only ones used by Starlings in 1976. The minor increase in
nests removed in 1975 (Fig. 1) is due to 1 site where boxes were checked every 3 to 4
days during May. As a result, 22 partially completed nests were removed from this area
alone.

Kessel (op. cit.) states that Starlings suffer a 50% annual mortality and that the
turnover period for a cohort to shrink to an insignificant portion of the population is 6
years. Assuming that the actual Starling population fell somewhere between the number
of boxes used and nests thrown out (Fig. 1), it is apparent that the major portion of the
Starling population was eliminated after 4 years of nest removal. Since reproduction
from boxes was prevented during the 1967-1973 period, an influx of birds from sur-
rounding areas was probably responsible for maintaining a small population.

Kessel ( pers. comm.) indicated that in New York during the mid-1940s, there were
always Starlings available from surrounding areas to take over vacated nest sites. We
did not determine what the population levels were on areas surrounding our study sites,
hut current Massachusetts Breeding Bird Atlas data indicate the Starling is the second
most common breeding bird in the state ( R. Forster, pers. comm.). The fact that
Starling production was eliminated on several of our study sites may have been due

GENERAL NOTES

289

Fig. 1. Number of Starling nests removed from Wood Duck boxes and number of
l)oxes used by Wood Ducks and by Starlings.

to the areas bolding isolated populations which, when removed, were not replaced.
However, we cannot offer any definite reason for this occurrence.

The Starling control program did not increase waterfowl production. As many ducks
used boxes in 1967 as in 1976 (Fig. 1). Duck production remained practically un-
changed on 4 sites, increased on 3, and decreased on 3. However, many other factors

290

THE WJLSON BULLETIN • VoL 90, No. 2, June 1978

also affected production during this period including water drawdowns, nest predation,
local hunting pressure, and vandalism. We believe that duck production would have
decreased during the 10-year study period had we not removed Starling nests since
Starling competition for boxes has been increasing since the early 1950s (Grice and
Rogers, The Wood Duck in Massachusetts, Mass. Div. Fish, and Wildl., 1965).

Nest removal appears to be an effective but time-consuming method of controlling
Starling populations in Wood Duck boxes. The use of Starling-deterrent nesting
cylinders described by McGilvrey and Uhler ( J. Wildl. Manage. 35(4) :793-797, 1971)
was evaluated by Heusmann et al. (Wildl. Soc. Bull. 5:14-18, I977j previously. GrabilTs
(Wildl. Soc. Bull. 5:69-70, 1977) use of Starling boxes attached to Wood Duck boxes
offers a third possibility for reducing competition for nesting sites between these species.

This study was a contribution of Massachusetts Federal Aid in Wildlife Restoration
Project W-42-R. — H W Heusmann and Robert Bellville. Massachusetts Division of
Fisheries and Wildlife, W'estboro 01581. Accepted 31 Oct. 1977.

Wilson Bull, 90(2), 1978, pp. 290-291

Unusual incubation behavior in Bobwhite. — A Bobwhite {Colinus virginianus)
nest with an incubating hen was found on 26 September 1975, 8 km west of Starkville,
Mississippi. The Bobwhite begins nesting in May in this area and this nesting effort
was perhaps the pair's 4th attempt. The nest and incubating hen were visited daily. On
3 October, when I attempted to photograph the hen on the nest, she left the nest giving
the broken-wing ruse. The hen ran about 35 m from the nest and was joined by a male
Bobwhite. The pair then flew off.

The nest contained 10 eggs, 1 of which proved to be infertile. One egg had successfully

Fig. 1. Partially hatched Bobwhite eggs with dead chicks.

(;enekal notes

291

hatched, l)ut the chick was not found. The remaining 8 eggs contained fully developed
hut dead chicks. The chicks had pipped and partially ringed their egg shells and then
became “entombed” (Fig. 1). Bohwhite chicks sometimes pip their egg shells but do
not hatch due to weakness and/or desiccation (Stoddard, The Bohwhite Quail, C.
Scribner’s Sons, NY, 1931; Rosene, The Bohwhite Quail, Rutgers Univ. Press, New
Brunswick, NJ, 1969; Stanford, Whirring Wings, Missouri Conservation Comm., Jefferson
City, MO, 1952).

Stoddard (op. cit. ) also said that Bohwhite have a strong nesting instinct and that an
incubating hen may remain lor a long period of time on eggs that do not hatch. The
nesting behavior reported herein is interesting because the hen remained on partially
hatched, dead chicks. Perhaps the stimulus of partially opened eggs kept the hen in-
cubating. Also, it has been thought that if only one egg hatches, the nesting instinct
would he satisfied ( Murray and Frye, The Bohwhite Quail and its Management in
Florida, Game Puhl. No. 2, Florida Game and Fish Comm., Tallahassee, 1964) . The 8
dead chicks had not started to decompose, but rather appeared to be desiccating. The
length of time the hen remained on the partially hatched eggs is unknown, hut the
hatching effort probably occurred several days prior to 3 October. — CEf)RGE A. Hurst,
Dept, of W^ildlife & Fisheries, Mississippi State Univ., Mississippi State 39762. Accepted
29 Dec. 1977.

Wilson Bull, 90(2), 1978, p. 291

A Cattle Egret-deer mutualism. — Cattle Egret (Bubulciis f^fs) -ungulate relation-
ships have been considered commensal ( Heatwole, Anim. Behav. 13:79-83, 1965; Jenni,
Fcol. Monogr. 39:245-270, 1969). Cattle Egrets feed more efficiently when taking in-
sects flushed by grazing cattle than when hunting apart from cows ( Dinsmore, Am. Midi.
Nat. 89:242-246, 1973). We observed Cattle Egrets feeding on 3 species of horse flies
(Diptera: Tabanidae) on the skin of the Virginia white-tailed deer (Odocoileus

virginianus) in the Okefenokee Swamp, Stephen Foster State Park, 30 km NE of Fargo,
Georgia from 28 ]\Iay to 2 June 1977. This is the first reported association between
Cattle Egrets and any North American nondomestic ungulate.

Observations (ca. 2.5 h) of several egrets and deer were made between 08:30 and
16:30 during the week. Commonly, 1 egret rode on the head or hack of a deer and
captured the large horse flies feeding on open wounds about the shoulder, head and face
of the deer. The deer appeared undisturbed by the presence or feeding activities of the
egrets. The species of horse flies observed (Tahanns petiolatus, Tahaniis americanus,
Diachlorus ferrngatus) inflict blood-letting wounds and may vector a variety of ungulate
microbial diseases occurring in the southeastern United States (James and Harwood.
Medical Entomology, Macmillan Puhl. Co., New York, 1969). Thus because both egret
and deer received benefits, the relationship was mutualistic. Whether the benefits are
significant to the success of the individuals is unknown.

We thank Drs. Roland R. Roth and Richard W. Rust for comments on the manuscript.
— Marc R. Halley and Wayne 1). Lord, Dept, of Entomology and Applied Ecology,
Univ. of Delaware, Newark, 19711. Accepted 3 Oct. 1977.

Wilson Bull, 90(2), 1978, pp. 291-295

A test of significance for Mayfield’s method of calculating nest success. — May-
field (Wilson Bull. 87:456-466, 1975) presented further suggestions to simplify his

292

THE WILSON BULLETIN • Vol. 90, No. 2, June 1978

previously published method (Mayfield, Wilson Bull. 73:255-261, 1961) for combining
all available nest data for a species to yield a less biased estimate of nest success than
n)igbt otherwise be obtained. In his more recent account he included a test of significance
that could be used for assessing the difference between estimates of success or, as in his
example, the difference between “mortality rates” at separate stages in the nesting cycle,
viz. between nest losses during incubation and nestling periods.

However, in simplifying his general methodology, I fear he may lead readers astray
with the test of significance that he proposes. Most statistical texts clearly state that for
a chi-square test, the elements in the contingency table must represent the frequencies
of discrete, mutually exclusive attributes. This is certainly not the case when “nest-days
with losses” are compared with “nest-days without losses” as in Mayfield’s Table 1.
Nest-days are arbitrary units, not discrete events. And any arbitrary time-interval, pro-
viding it is not less than the actual time required for the predation or destruction of a
nest, can never have a frequency greater than the actual number of nests lost. Time-
intervals without losses are, however, a function of the total time the nest was active.

Mayfield was aware of this problem. After finding that the test yielded a non-
significant result, he stated: “. . . the difference may still be real. A larger sample might
increase the confidence level. Changing to a smaller unit [my italics] — say, nest-hours —
might seem to hold out promise of a different result, since it multiplies the sample size
without enlarging the losses. . . .” This very suggestion is a clear violation of assump-
tions underlying the statistical test. Normally, subdivision of the data would increase
the calculated significance of the test, because the data in each class should be in
equivalent units, e.g., failures versus successes. The caution against using chi-square
tests on percentage data, echoed by so many texts (e.g., see discussions in Goldstein,
Biostatistics, Macmillan, New York, 1964:102; Snedecor and Cocbran, Statistical Methods,
6th ed, Iowa State Univ. Press, Ames, 1967:21; Vann, Fundamentals of Biostatistics,
Heath and Co., 1972:156; Woolf, Principles of Biometry, Van Nostrand, Princeton, N.J.,
1968:235), is a warning against an analogous procedure. If the total size of a sample
were less than 100, then the use of percentages would artificially inflate the values in
the contingency table and the analysis would indicate too great a significance. On the
other hand, if the sample were greater than 100, the use of percentages would decrease
the indicated probability that the difference is significant. Vann (op. cit.:164) indicates
that to double the sample size in each cell of a contingency table is to double the chi-
square value that will be calculated from it.

By suggesting, as does Mayfield, that “nest-days without losses” be increased by
changing to hours [why not minutes? seconds?], while “nest-days with losses” cannot
be changed because they are clearly fixed by nest losses, is to reduce the sensitivity of
the test. As the denominator of each proportion is increased in this arbitrary manner,
the proportions will both tend toward zero. It can be shown that when the sample
size in only one column of a contingency table is increased, the calculated value of chi-
square will increase; but, because tbe proportions are changing, as the sample size
becomes infinitely large, chi-square approaches a finite value. This is quite unlike the
situation when the sample size is increased in both columns of the contingency table,
for then the proportions are unchanged and as the sample size becomes infinitely large,
chi-square approaches infinity as it should. Thus the test proposed by Mayfield is too
insensitive to differences between such proportions. In fact, the = F16 calculated
by Mayfield cannot be increased beyond x" = 1.20, regardless of the increase in data
to the column he suggests.

There seems to be no ready remedy. Although Mayfield’s method might yield a less-

GENERAL NOTES

293

Comparison

Table 1

OF Mortality During Incuration
Kirtland’s Warbler ( Dendroica

AND Nestling
KJRTLANDII)

Periods of

Average

Calculated

Total

mortality rate

no. of nests

nests

for period*

lostf

succeeded

found]:

Incubation period

0.44

67.8

86.2

154

Nestling period

0.24

34.6

109.4

144

Totals

102.4

195.6

298

x"= 13.20; df- 1

; P< 0.0005

* M = 1 — S, where S is the average survival rate estimated by Mayfield (1975) for the in-
cubation or nestling period: S = (1 — P)“ and P = 0.040 and 0.030, the mortality rate per nest-

day; n = 14 and 9, the incubation and nestling periods in days.

t Calculated by multiplying ‘total nests found’ by ‘average mortality rate for period.’ Number
of nests succeeded can then be obtained by subtraction,
t Obtained from Mayfield (1960).

biased estimate of nest failure than obtained by ignoring nests other than those found
before eggs had been laid, the fact remains that the only independent units available
for statistical analysis are not nest-days or nest-hours but nests. The mortality rate
that his method attempts to estimate is that which would be obtained if all nests could
have been found and followed through the season, viz. failed nests/all nests. Thus,
the only reasonable units for the marginal totals of the contingency table are the actual
numbers of nests involved. I suggest calculating the frequencies within the table from
the estimate obtained by Mayfield’s method; this ensures the retaining of the accuracy
of his method, while using only the total number of nests in the chi-square analysis
ensures the satisfying of an important underlying assumption. This assumption is set
out unequivocally by Siegel ( Nonparametric Statistics for the Behavioral Sciences,
McGraw-Hill, New York, 1956:44) : “To use in testing a hypothesis . . . cast each
observation into one of k cells. The total number of such observations should be N, the
number of cases in your sample. That is, each observation must be independent of every
other; thus one may not make several observations on the same person and count each
as independent [read nest for person]. To do so produces an inflated ‘A’.” And on
p. 109: “A is the sum of each group of marginal totals. It represents the total number
of independent observations. Inflated A’s invalidate the test.”

The test I suggest is illustrated with Mayfield’s data in Table 1. Basically it can be
seen to be an extension of Mayfield’s estimate of nest success itself. As he points out,
to ignore incomplete records of nests is to bias results towards understating mortality
and overstating success. He argues that his method brings mortality estimates to a more
realistic level. My test makes use of these presumably more accurate estimates of
mortality and survival to calculate the actual number of nests that should have failed
or succeeded in a given sample of nests found at all stages. The data used by Mayfield
in his Table 1 were based on 154 nests in the incubation period and 144 in the nestling
period (Mayfield, The Kirtland’s Warbler, Cranbrook Inst. Sci., 1960:193, 198). From
the estimate that he calculated, the actual number of nests lost in the incubation period
must have been closer to 67.8 than the 35 that he observed. Likewise, losses during the
nestling period must have been closer to 34.6 than to the 22 he noted. The chi-square
test in Table 1 shows that the difference between mortality rate during incubation and

294

THE WILSON BULLETIN • Vol. 90, No. 2

nestling periods is in fact highly significant ( = 13.20; P <1 0.0005), not non-
significant as Mayfield concluded. Thus there would seem little justification for lumping
the 2 stages to obtain a single rate. Mayfield stressed that the observer should state
reasons for electing to lump or not to lump data from these 2 sources. The method that I
have described provides an objective test of the homogeneity of the samples and thus a
criterion to aid the decision to lump or not.

It may concern some readers that my test does not make use of “observed” and “ex-
pected” values as usually seen in a standard contingency table to he analysed by chi-
square. They may prefer to think of the observations being made indirectly through
Mayfield’s method; the marginal totals are in fact observed. However, it is important to
realize that these values are established prior to and quite separately from the “ex-
pected” values in the standard chi-square analysis, and from the point of view of that
analysis are indeed “observed” values.

Doubtless, more sophisticated analyses could be developed for data of this kind, but
I think that chi-square still provides the easiest test and one that fits nicely with May-
field’s desire to simplify the procedure as much as possible. As presented here, it has
the advantage of being no more biased by sample size than Mayfield’s estimate of nest
success might he, for the cell frequencies in the contingency table are based on that
estimate. A possible source of bias, in some tests, lies in the fact that many nests from
the nestling period are the same nests as those included in the incubation period, i.e.,
the rows of Table 1 may not l)e entirely independent. However, this will probably be of
little importance and, in any case, will tend to make the test more conservative. Al-
though not discussed by Mayfield, Yates’ correction for continuity (see any of the above
references to statistical works) is sometimes applied to such 2x2 contingency tables.
But retention of the decimal part of the estimated numbers of nests will render this
correction less important, although it will change the calculated chi-square value con-
siderably if the sample size is small.

The method can be extended readily to tables with more than 2 rows and to other
situations where tests of homogeneity are appropriate. The hypotheses under test, in
any event, relate to differences between or among the estimates for incubation, nestling,
or other periods obtained by Mayfield’s method.

Readers using Mayfield’s method should appreciate that the nest-day selected by him
as the unit for analysis, although perhaps the most defensible on biological grounds
because it provides a complete cycle for all possible predation, is an arbitrary choice.
Mayfield does not discuss the selection of this unit, hut his implication 11975 op. cit.)
that nest-hours might be useful suggests that he is unaware of the difficulty. To change
the unit of time also changes the success estimated for the overall period. For example,
had his calculation of nest survival during incubation ( from 35 nests lost in 878 nest-
days of observation; daily mortality P — 0.040) been based on nest-minutes, he w'ould
have concluded that 57.23% rather than 56.58% of nests succeeded. At the other ex-
treme, nest-fortnights <14 days is after all the incubation period of the species in his
example) would yield a success rate of 44.19%. Other than through faith in the im-
portance of circadian periodicity, it seems difficult to justify why the nest-day should
yield the success rate closest to the actual one.

Although Mayfield devised his method ostensibly to reduce the bias of the conven-
tional one. Green (Wilson Bull. 89:173-175, 1977) has shown that if a breeding popu-
lation contains birds that can he categorized according to the vulnerability of their
nests, Mayfield’s method will yield a biased result. Thus, when nests can he separated
on any basis le.g., habitat difference, age of owner, height, time in the season, first or

GENERAL NOTES

295

second clutch), it would appear advisable to use Green’s modification of Mayfield’s
method. When such separation of data cannot he made, the dilemma of the researcher
will be in not knowing whether the bias of Mayfield’s method towards understating
success is greater or less than the bias of the conventional method in overstating it. —
Douglas D. Dow, Dept, of Zoology, Univ. of Queensland, Brisbane, Australia, 4067. Ac-
cepted 20 Jan. 1977.

Wilson Bull, 90(2), 1978, pp. 295-297

Mirror image versus conspecific stimulation in adult male Zebra Finches.- —

Mirror image stimulation (M.I.S.) has been used to study social responses in a variety
of animals (Kaufman and Hinde, Anim. Behav. 9:197-204. 1961; Svendsen and Armitage,
Ecology 54:623-627, 1973). The advantage of this technique over direct visual contact
with a conspecific is control of the stimulus by the experimenter. While this procedure
has been conducted with several species, only chimps have so far shown the ability to
recognize their own image (Gallup, Science 167:86-87, 1970).

Siamese fighting fish iBetta splendens ) exhibit a preference for M.I.S. over visual ac-
cess to a conspecific ( Baenninger, Psychon. Sci. 4:241-242, 1966). This phenomenon
also has been shown to exist in House Sparrows i Passer domesticus) and parakeets
i Melopsittacus undulatus) (Gallup and Capper, Anim. Behav. 18:621-624, 1970). I ex-
amined this phenomenon in Zebra Finches iPoephilo guttata).

Ten adult male Zebra Finches were tested in a continuous choice situation of M.I.S.
and visual access to a conspecific. The testing apparatus was modified after the plan
of Gallup and Capper (op. cit.). It consisted of a plywood box 72 cm long, 42 cm
wide and 38 cm high. The ceiling had an opening 43 cm long and 26 cm wide covered
by fine screen.

Two perches, 17 cm long, were placed 8 cm above the floor and centrally located on
each side of the cage. One perch was associated with an 8 X 5 cm mirror; the opposite
perch was placed in front of a plexiglass window of the same size. This window allowed
visual access to a 14 cm long, 14 cm wide and 18 cm high target cage. The target cage
contained one adult male conspecific. A cardboard partition in the test cage kept the
test bird from seeing the stimulus associated with the opposite perch. A continuous
supply of food and water was available in the target cage and on each side of the ex-
perimental cage.

Testing was initiated by placing one l)ird in the experimental cage and a conspecific
in the target cage. A photoperiod of 12 h was controlled by a light source placed 40 cm
above the testing apparatus. Time spent on each perch was electrically quantified.
The weight of the bird on the perch triggered a microswitch hooked up to an electric
timer. Timers were reset at the end of each 24 h period; this procedure was continued for
a period of 9 days.

Figure 1 shows the amount of time spent on the perch associated with the
mirror and the amount of time spent on the perch associated with the target cage.
Day 1 was not included in the graph as this was considered an adjustment period. The
data show an overwhelming preference for interaction with the conspecific compared to
the almost total absence of time spent on the perch associated with the mirror. The
preference for conspecific visual access continued until day 6 after which preference
for M.I.S. took place. Preference for M.I.S. continued through the final 3 days of testing.

All of the 10 individuals tested showed an overwhelming initial preference for visual
access to the conspecific and later exhibited a preference for M.I.S. which continued

296

THE WILSON BULLETIN • Vol. 90, No. 2, June 1978

Fig. L Mean time in minutes spent in front of a mirror (solid circles) and in front
of a conspecific (open circles) by 10 adult male Zebra Finches as a function of time.

throughout the testing period. The earliest shift in preference took place on day 4 by 1
individual, and the latest shift was exhibited on day 8 by 2 individuals. One individual
shifted preference on day 5, 2 on day 6 and 4 on day 7. Due to the abbreviated period
of testing it was not possible to determine whether this shift was more than a temporary
phenomenon.

The fact that after a period of time there is a definite preference for M.I.S. in adult
male Zebra Finches suggests this phenomenon may occur among other bird species.
The preference for M.I.S. was shown in the absence of auditory cues. This tends to
support the hypothesis of Gallup and Capper (op. cit.) that the mirror image is being
perceived as a supernormal stimulus. Superficially, there may seem to be little difference
between the mirror image and the conspecific, but the mirror image will always be
both predictable and compatible with the animal’s behavior. The mirror image may
also be considered a novel stimulus. It has been demonstrated with guppies (Lebistes
reticulata) (Russell, Anim. Behav. 15:586-594, 1967) that novel stimuli may be investi-
gated less in a strange environment. This may explain the initial lack of interaction with
the mirror image. The fact that the animal did not habituate to the image may be due
to the al)breviated duration of the experiment. The complexity of a stimulus increases
its novelty ( Berlyne, Conflict, Arousal and Curiosity, McGraw-Hill, New York, 1960:43).

(;eneral notes

297

This may he true of the mirror image and at least partly responsible for the lack of
habituation. Future work considering the j)reference for M.l.S. over an extended period
of time with a variety of species may give some idea as to the ada])tive significance of
this behavior.

I would like to thank Roger J. Raimist for his helpful suggestions during study.
Cindy Banas made the graph. This research was partially funded by the Student Re-
search Committee, Life Science Dept., Glasshoro State College. — Michael J. Ryan, Life
Science Dept., Glasshoro State College, Glasshoro, NJ 08028. < Present Address; Dept,
of Zoology, Rutgers IJniv., Newark, NJ 07102). Accepted 30 Mar. 1977.

Wilson Bull, 90(2), 1978, p. 297

Protocalliphora infestation in Great Horned Owls.- On 5 May 1977, 6.4 km
southwest of Foley, Minnesota, I collected several dipteran larvae from the ear cavities

of a nestling Great Horned Owl, Buho virginianus. I raised the larvae to adult flies,

which were identified by Dr. Curtis W. Sahrosky, Systematic Entomology Laboratory,
U.S. National Museum as Protocalliphora avium Shannon and Dohroscky. The adult

flies resemble blue-bottle flies, but lielong to the family Calliphoridae; the blow flies.

Protocalliphora have been found to parasitize a number of raptors, including Long-eared
Owls, Asia otus (Shannon and Dohroscky, J. Washington Acad. Sci. 14:247-253, 1924),
and Red-tailed Hawks, Buteo jamaicensis. Red-shouldered Hawks, Buteo lineatus, and
Cooper’s Hawks, Accipiter cooperii (Sargent, Auk 55:82-84, 1938). I found dipteran
infestations to be quite common in Great Horned Owl and Red-tailed HawT nestlings in
central Minnesota. During the past 2 years, 41 of 73 ( 56% ) Red-tailed Hawks, and
25 of 46 (54%) Great Horned Owls that I banded were infested. The larvae were
located in the ear cavities of most nestlings although some Avere found in the nape area.
Unfortunately, only from the one nest did I have the dipterans precisely identified.
Other raptors may also be afflicted with this parasite, but there appear to be few docu-
mented cases. Protocalliphora eggs are apparently deposited in the nest debris; the
larvae suck blood intermittently for 14—20 days and pupate for about 10 days before the
adult flies emerge (Coutant, J. Parasitol. 1:135-150, 1915). The blood-sucking larvae
usually do not seriously harm large species; however, they may weaken, kill, or force
smaller passeriformes from their nests (Johnson, Ann. Entomol. Soc. Am. 22:131-135,
1929). — Robert T. Boiim, 520 7th Ave. North, Sauk Rapids, MN 56379. Accepted 21
Feh. 1978.

Wilson Bull, 90(2), 1978, pp. 297-299

Territorial defense of a iieetar source by a Palm \'i'arhler. — Territorial defense
of nectar has been documented in several species of wintering parulids, for example:
Cape May Warbler, Dendroica tigrina (Kale, Auk 84:120-121, 1967; Enden, Wilson
Bull. 85:71-74, 1973 ), Palm Warbler, D. palmarum ( Emlen, op. cit.), and Yellow-
rumped Warbler, D. coronata (Woolfenden, Auk 79:713-714, 1962). It is the purpose
of this note to document further the defense of a nectar source by a Palm Warliler and
the disproportionate amount of time it spent chasing conspecifics from flowers as com-
pared with the time sjient chasing 2 other parulid species.

The following observations were made over a 5 h period (07:00 12:18) in Bayside
Park, Miami, Florida on 9 March 1975. An unl)anded Palm Warbler was observed

298

rilE WILSON MULLETIN • VoL 90, No. 2, June 1978

Table 1

SUMMAJ{Y OF THE TiME Sl'ENT IN PUHSUIT BY A PaI.M WabBLEB DEFENDING ElOWEBS OF A

Tigeb’s Claw Tbee

Species

\iimher

pursuits

Average tiine
of pursuit
(sec)

Standard

error

Range

(sec)

Northern Parula

1.4

25-61

Yellow-rumped Warbler

1.5

32-60

Palm Warbler

218

2.2

61-321

constantly as it defended the flowers of a Tiger’s Claw Tree i Erythrina sp.). The leaf-
less tree was 10-12 m tall with a crown diameter of approximately 10 m. The following
species regularly visited the flowers on the tree hut were not chased by the Palm
Warbler: Common Flicker (Colaptes auratus) , Red-bellied Woodpecker (Melanerpes

carolinus) , Fish Crow iCorvus ossifragus) , Mockingbird iMimus polyglottos) , Starling
(Sturnus vulgaris), and Spot-breasted Oriole {Icterus pectoralis) . The Palm Warbler
appeared to he actively defending the flower-covered tree from Northern Parulas {Parula
americana) , Yellow-rumped Warblers, and other Palm Warblers. Only parulids were
chased from the tree. All chases were timed with a stop watch; only pursuits which I
could see from start to finish are listed in Table 1.

The Palm Warbler seemed to he feeding on nectar during the observation period.
For short periods of time (10-15 min) this individual stayed high in the tree calling and
flicking its tail. From this location it often sighted and chased intruders.

From 09:00 to 09:30 this individual spent about 60% of the time sitting on a look-
out perch, 30% chasing intruding parulids, and 10% feeding at flowers. This time
budget is similar to that noted by Emlen ( op. cit. ) for a wintering Cape May Warbler
defending a nectar source in the Bahamas.

My observations (Table 1) show that the Palm Warbler spent more time chasing con-
sjiecifics (average 218 sec per chase) from the tree than either species of parulid (average
of 39 sec for parulas and 42 sec for Yellow-rumped Warblers). Compaiisons of the
average pursuit times for Palm with Northern Parulas and Palm with Yellow-rumped
warblers both showed significant differences ( p < .05 ) using a t-test. All 3 species
fled from the territorial individual in the same manner and it is unlikely that the dif-
ferences in pursuit times are due to differences in the behavior of the fleeing individuals.

Why more time and energy should he expended in the pursuit of conspecifics is not
altogether clear. It is unlikely that a conspecific is more of a threat in nectar consumption
than a parula or Yellow-rumped warbler. All 3 parulids appeared to feed in the same
manner and presumably removed equal amounts of nectar. During 11 lengthy pursuits
of conspecifics both Yellow-rumped and parula warblers flew into the unguarded
tree and fed until chased by the returning territorial bird. The disproportionate amount
of time spent in pursuit of conspecifics left the nectar source unguarded and thus
available to other parulids. It appears that shorter pursuits of conspecifics would have
been more efficient in guarding the tree from competing parulids.

The selective pressures for species recognition are probably quite strong and the
appearance of a conspecific at a defended food source arouses a stronger aggressive

GENERAL NOTES

299

response than the appearance of other species. By responding more strongly to con-
specifics, a territorial individual might leave the nectar source unguarded and thus
available to other competitors. The rarity with which such nectar sources are en-
countered and the more unlikely condition that another larger species is not already
defending it, give little time for selection to “finely-tune” this defense behavior. Nectar
specialists, i.e. hummingbirds and sunbirds, tend to chase all competitors of the same size
or smaller with equal vigor. The Palm Warbler which on rare occasions takes nectar
might respond to a competitor with which it is most familiar. Aggression towards con-
specifics is a common behavior found within the repertoire of possible responses and
may account for the differential pursuit times.

Selection should favor defense of a feeding territory as long as the energy gained from
exclusive use of that defended nectar source is greater than the energy expended in its
defense (see Stiles and Wolf, Auk 87:467-491, 1970; Wolf, Condor 72:1-14, 1970).
Generally interspecific dominance is based upon size, for larger species are either dif-
ficult or impossible to drive out of the territory. This territorial Palm Warbler ignored
all intruding larger species and chased only parulid species of equal size.

1 am grateful to K. WL Corbin, J. H. Rappole, P. J. Regal, and H. B. Tordoff for
reviewing this manuscript. — Joseph M. Wunuerle, Jr., Dept, of Ecology and Behavioral
Biology., Bell Museum of Natural History, Univ. of Minnesota, Minneapolis 55455. Ac-
cepted 22 Feb. 1977.

Wilson Bull, 90(2), 1978, pp. 299-301

Ring-billed Gull pair with 2 nests. — On 13 May 1975, I observed 2 nests of Earns
delawarensis in the Calcite Colony (Rogers City, Presque Isle Co., Mich.) that were
unusually close together. The nests had a common rim on one side and both contained
eggs. Observations verified that only 2 gulls, apparently a pair, were attending the
double nest. The nests were in a i)ortion of the colony that has been used repeatedly
since at least 1958.

The largest and most complete nest ( i.e. that with the most nest material I contained
4 eggs while the other had 2 (Fig. 1). Two adult-plumaged gulls (i.e. lacked terminal
tail bands and other characters normally indicative of birds less than 3-years old; see
Ryder, Wilson Bull. 87:534, 1975) attended the nests. Each of the gulls consistently in-
cubated the same clutch of eggs during my two 8-h observation ])eriods before color-
marking. Neither of the birds left the nest site while I was present on these 2 days.
This represents unusually long incubation bouts for each bird as the mean duration of
shifts for pair members at this colony is 1.8 h (.Southern, pers. observ. ).

The incubating birds were tolerant of one another and body contact was not unusual.
Fre(}uently the head of one bird touched the wing, tail, or back of the other. Occasionally
their bodies were aligned i)arallel to one another facing in the same direction. The 2 gulls,
either singly or in comI)ination, threatened incu!)at.'ng neighbors and territory intruders.
Botli birds arranged nesting material, including that in the common wall l)etween the
nests.

Once the gull attending the 2-egg clutch left the nest and stood unchallenged near
tlie adjoining nest. Before returning to the nest, it chased an intruding neighbor from
the territory. While standing at the nest before settling, it again threatened the neighbor
(with ()j)en-bill thrusts), this time in unison with its partner on the adjacent nest. During
my observations no other gulls approached the double nest without being challenged by
one or both attending gulls.

300

THE WILSON BULLETIN • VoL 90, No. 2, June 1978

Fig. 1. Doulile Ring-billed Gull nest tended by 2 adult-plumaged birds.

To determine if more than 2 Ring-bills were involved, the incubating birds were color-
marked on the 3rd day of observation. This was accomplished liy placing a paint-soaked
swab of cloth on the nest rim so that any gull settling to incubate was marked on some
part of its ventral plumage. Yellow' was used at the 4-egg nest and blue at the other.
Both gulls appeared somewhat alarmed by the presence of paint on the nest as they
had resettled quickly following my previous daily check on nest contents. Following
marking, the “yellow gull” rearranged nesting material as if attempting to cover up the
spot. This action stimulated the other gull to contribute nest material and both worked
on the spot. The gull attending the 2-egg clutch was hesitant to settle on the blue paint
and after doing so immediately left the nest and stood nearby. Within 2 min it returned
with nest material and wmrked on the painted area. About 10 min later both birds
settled on their respective nests but the “blue gull” continued to rise and resettle.
About 3.5 h later the blue gull ceased sitting on the nest and remained nearby with its
plumage sleeked. It continued to defend the nest and eggs against intruders.

The following morning the yellow gull was standing near the 4-egg nest that now
was being incubated by the blue gull. The 2-egg clutch had been abandoned and was
not incubated thereafter. Four days later the 2 eggs were gone and the nest material,
or at least that spotted by blue paint, had been incorporated into the remaining nest.
The 2 gulls shared incubation duties thereafter at this nest. Four young eventually
hatched and w^ere cared for by both adults. Continuous observations were not made of
the family group following hatching and so the role played by each adult in caring
for the young is unknown.

Since the birds were not sexed it remains possible that the 2 gulls w'ere other than a
mated male and female < e.g. 2 females). Polygyny has been reported in other gulls
(e.g. Lams ar^entatiis. Shugart and Southern, in press) but in this particular case a 3rd ;|

GENERAL NOTES

301

gull was not involved, or at least did not share incubation duties. Therefore, this may he
the first recorded instance of a pair of Ring-billed Gulls, and possibly any gull species,
having 2 clutches of eggs simultaneously. — William E. Southern, Dept, of Biological
Sciences, Northern Illinois Univ., DeKalb 60115. Accepted 22 Feb. 1977.

Wilson Bull, 90(2), 1978, pp. 301-302

Clutch size and nest placement of the Pied-hilled Grebe in Manitoba. — The

Pied-billed Grebe { Fodilymbus podiceps) breeds widely throughout North, Central, and
South America ( Palmer, Handbook of North American Birds, Vol. 1, Yale Univ. Press,
New Haven, Conn., 1962). Despite this, relatively few studies of its breeding biology
have been conducted. I obtained information at 2-4 day intervals between 19 May and
25 July 1973 on nest placement and clutch size of the Pied-hilled Grebe in a prairie pothole
area south of Minnedosa, Manitoba. This area has been described in detail by Ferguson
< M.Sc. thesis, Univ. Manitoba, Winnipeg, 1977) and more generally by Ehrlich et al.
(Man. Soil Surv. Rept. No. 6. 1957), Bird (Canada Dept. Agric., Contr. No. 27. 1961),
and Keil et al. (Can. Wildl. Serv. Rept. Ser. No. 18. 1972).

Semi-drought conditions existed in the study area in 1973. The ephemeral and
seasonal potholes ( following Stewart and Kantrud, Resource Puhl. 12, Bur. of Sport
Fish, and Wildl., 1971 ) were dry or nearly so. The water levels in most semi-permanent
and some permanent potholes were low, in some instances exposing entire stands of
emergent vegetation.

In all cases (N=:53) only 1 pair of Pied-hilled Grebes nested per pothole; this
was also the case with most Horned Grebes i Podiceps auritus) nesting in the Minne-
dosa area (Ferguson 1977) and the Horned and Pied-hilled grebes in North Dakota
(Faahorg, Wilson Bull. 88:390-399, 1976).

Of 42 nests found with eggs in 1973, 29 failed; most of the failures were washed out
before the clutches were completed. The average size of 22 completed first clutches was
6.8 eggs (1 of 5 eggs, 3 of 6, 17 of 7, 1 of 8). The mean size of 6 suspected replacement
clutches was 7.2 eggs ( 5 of 7 eggs, 1 of 8). Glover (Wilson Bull. 65:32-39, 1953) also
found no difference in size between first and first replacement clutches ( see also Miller,
Cassinia 32:22-34, 1943) and lumped them in his calculations. Palmer (1962) in-
dicated that Pied-hilled Grebe clutch sizes do not decrease as the season progresses.
Ferguson (1977), however, noted such a decline in Horned Grebe clutch size in the
Minnedosa area. The clutches used to calculate clutch size in my study were all
initiated within a 2-week period from late May to early June.

I obtained additional information on Pied-hilled Grebe clutch sizes in Manitoba,
Saskatchewan, and Alberta from the Prairie Nest Records Scheme (PNRS) and for
British Columbia from the British Columbia Nest Records Scheme (BCNRS). Nests
where the clutch size was the same after 2 visits, 2-3 days apart, were used. The mean
size of 6 such clutches from the PNRS was 6.7 eggs (extremes, 3 and 10). In 19 other
nests visited only once by the observer, there were 5 with 6 eggs, 6 with 7, 5 with 8 and
3 with 9 ( X = 7.3) .

There is no apparent increase in clutch size from south-to-north in temperate North
America (Table 1). The small clutches found in Idaho are not explained. The small
sample of clutch sizes from Central America (Costa Rica, Honduras) suggests an overall
latitudinal increase in clutch size. However, much more work needs to he done there on
Pied-hilled Grebes. In the Atitlan Grebe (P. gigas) of Guatemala, LaBastile (Wildl.
Monogr. No. 37, 1974) reported a mean clutch size of 2.9 eggs, with extremes of 2 and 5.

302

THE WILSON HULLETJN • Vol. 90, No. 2, June 1978

Clutch

Table 1

Size of the Pied-Billed Grebe

Locality

No. Nest.s

Mean Clutch
Size (Range)

Source

Costa Rica

(3-4r

F. C. Stiles {in litt.)

Honduras

Monroe ( Ornithol. Monogr. No. 7,

Louisiana

67"

7.0 (5-9)

1968)

Chabreck (Auk 80:447-452, 1963)

Pennsylvania

6 or 7 (5-9)

Miller (op. cit., 1943)

Iowa

97c

6.2

Iowa

41“

(2-10)

4.3

6.1 (4-8)

Glover (op. cit., 1953)

Provost (Am. Midi. Nat. 38:483-

Idaho

4.3

503, 1947)

Wolf (J. Wildl. Manage. 19:13-

North Dakota

6.7 (4-11)

23, 1955)

Stewart ( Breeding Birds of North

Manitoba

6.8 (5-8)

Dakota, Lund Press, Minne-
apolis, 1975)

This study

British Columbia

7.0 (5-9)

BCNRS

Alberta, Sask., Man.

6.7 (3-10)

PNRS

“ It was not known whether the clutch of 3 was complete (F, G. Stiles, in lift.).

In salt marshes.

^ Successful clutches.

Unsuccessful clutches.

Water depth measured at 31 Pied-billed Grebe nests averaged 35.5 cm. No nests
were located in water less than 12.7 cm deep; 4 nests were in water 14-24 cm deep.
A nest discovered on 28 June 1973, which contained 6 eggs (3 of them pipped), was in
14 cm of water, 0.3 m from shore. To leave the nest, the adult skittered on the water to
open, deeper water 2.1 m away. The mean distance of 25 nests from shore was 5.3 ± 2.4
m and from open water, 1.3 ± 0.9 m. The dominant species of emergent cover at 53
nests was bulrush i Scirpus spp.; 37 nests, 69.8%), cattail (Typha lati folia; 15 nests,
28.3%), and whitetop {Scholochloa festucacea; 1 nest, 1.9%).

I thank R. W. Campbell and H. W. R. Copland for providing information on Pied-
billed Grebe nests, contained in the BCNRS and PNRS, respectively. J. Faaborg and
R. W. Storer commented on an earlier draft of the manuscript. A. LaBastille, E. R.
Blake, P. Skid, and F. C. Stiles kindly responded to my inquiry regarding clutch sizes
of Pied-billed Grebes in Central America. Funds were provided by the National Re-
search Council of Canada ( A9556 ) and the University of Manitoba Research Board.
This is contribution number 35 of the University of Manitoba Field Station (Delta
.Marsh).- SpENCKit C. Sem.y, Dept, of Zoo!., Univ. of Manitoba, ffinnipeg, .Man.
R3T 2N2. Accepted 10 Nov. 1977.

GENERAL NOTES

303

Wilson Bull., 90(2), 1978, p. 303

Nest placement in Sage Thrashers. — Nest site selection and placement has been
shown to be important to nest success in various avian species ( Goddard and Board,
Wilson Bull. 79:283-289, 1967; Tenaza, Condor 73:81-92, 1971). Birds nesting in
desert environments are particularly adapted to construct, place, and orient nests so as
to maintain a favorable beat balance in eggs, young, and adults (Ricklefs and Hains-
worth. Condor 71:32-37, 1969; Austin, Auk 93:245-262, 1976).

During the summer of 1976 I studied nest placement in Sage Thrashers {Oreoscoptes
montanus) on the Snake River plain north of Pocatello, Idaho. The habitat is dominated
by sagebrush {Artemisia tridentata) . Of 15 first nests built and hatching young over
the period 7 May to 29 May 1976, 13 were placed on the ground beneath the sage and
close to the trunk. The average nest height was 2.9 cm from the bottom of the nest
to the ground. Of 6 nests built and hatching young over the period 24 June to 14 July,
5 were placed off the ground in the branches of the sage. The average height of these
nests was 22.7 cm. The difference in heights is significant, 0.005 <7 P (U ^ 79) <[0.01
(Mann- Whitney U test, Siegel, Nonparametric Statistics for the Behavioral Sciences,
McGraw-Hill Book Co., Inc., New York, 1956). A typical nest is about 10 cm from top
to bottom.

This difference in placement of first and second nests could be due to a number
of factors. 1 do not know whether second nests were actually second broods of particular
pairs or perhaps the nests of later nesting birds. The location of the later nests near
first nests suggests the former. I believe the placement was a result of adjustment to
avoid thermal stress. In May 1977 temperatures were recorded at the mean nest heights
of 2.9 cm and 22.7 cm on 5 days, each day at a different location in the sagebrush. This
mean maximum temperature at the upper position was 29.3°C and that at the lower
32.3° C. The mean minimum temperature at the upper position was -7.8° C and that at
the lower -3.0° C.

Thus, first nests built early in the year and placed on the ground would benefit from
the warmer temperature at ground level. Second nests placed higher in the sage and
away from the main stem would benefit from air circulation and convective heat loss
to the cooler air. This has been shown an effective means of heat dissipation from nests
of the Cactus Wren (Campylorhynchiis briinneicapillum) as reported by Ricklefs and
Hainsworth (op. cit.).

Twelve of 15 first nests were built directly under a thick overhanging branch of sage,
and another was placed under an old nest from some previous year. This would serve
to reduce heat loss to the night sky as reported for hummingbirds (Calder, Condor
73:314-321, 1971). No second nests were built under thick overhanging structures.

The depth of nest cups, as measured after the fledging of young, was greater in
second nests <8.2 cm) than in first nests <4.7 cm) though not significantly so,
0.10 < P (U ^ 17.5 ) <0.20. 1 suggest that the deeper cups would protect the eggs

and young from direct exposure to the sun while adults were off the nest.

No significant relationships were found with respect to orientation of nests from the
main body of the sage.

This study was made while conducting research sponsored by the Frank M. Chapman
Memorial Fund of the American Museum of Natural History and the Department of
Biology, Idaho State University. — Tekrki.l I). G. Rich, Dept, of Biology, Idaho State Univ.,
Pocatello 83209. Accepted 28 Apr. 1977.

304

THE WILSON BULLETIN • Vul. 90, No. 2, June 1978

Wilson Bull., 90(2), 1978, pp. 304-305

Great Black-hacked Gulls breeding in salt marsh in New Jersey. — Great Black-
hacked Gulls [Larus nwriniis) have recently expanded in the northeastern U.S., nesting
along the east coast as far south as Jamaica Bay, New York ( Feakall, Kingbird
17:69-73, 1967), where they normally nest on sandy islands, shingles, moors, and grassy
areas. In 1972, Parnell and Soots (Auk 92:154-157, 1975) found several adults and one
nest on a dredge island in North Carolina. This report documents the invasion and
successful nesting of Great Black-hacked Gulls in the salt marshes of New Jersey. Since
breeding behavior and success data are usually unavailable from the first pairs nesting
in an area, I present these data.

In 1976, Great Black-hacked Gulls nested on 5 salt marsh islands in Ocean County,
New Jersey ( 39°41'-39°46' N ) : 8 pairs on Clam Island, 2 pairs each on Sloop, Caravel,
and Egg islands, and 1 pair on Sandy Island. Spartina patens and S. alterniflora
dominate these low salt marsh islands with Iva Jrutescens growing in the higher areas.
I checked nests on Clam Island 2 to 4 times a week from early April until late July,
on Caravel Island weekly, and on the other islands once or twice a month.

On all islands Black-hacked Gulls nested within Herring Gull (L. argentatus)
colonies of 15 to 800 pairs. Laughing Gulls (L. atricilla) and Common Terns i Sterna
hirundo) nested on Clam Island and Egg Island. Common Terns and Black Skimmers
{Rynchops nigra) nested on Caravel Island.

Black-hacked Gulls generally nested under Iva hushes on grass in the center of the
densest area of Herring Gulls. Channels divide Clam Island into several suhislands, each
containing one or more small areas with Iva hushes. On 9 April when I first visited
Clam Island, I found 16 Herring Gull nest scrapes and 2 partially completed Black-
hacked Gull nests on the NE subisland, 10 Herring Gull nest scrapes and 2 completed
Black-hacked Gull nests on the NW suhisland, and 8 Herring Gull scrapes and 2 Black-
hacked nests on the SE suhisland. Although nesting activity began on all suhislands at
the same time, the Black-hacked Gulls were separated by 500 to 800 m. Thus, Black-
hacked Gulls spaced themselves as if solitary with respect to conspecifics, hut colonially
with respect to Herring Gulls.

In all colonies Black-hacked Gulls initiated egg-laying between 7 and 18 April, during
the earliest period of egg laying for Herring Gulls. All nests found contained 3 eggs.
Erwin (Wilson Bull. 83:152-158, 1971) found that Black-hacked Gulls lay eggs earlier
than Herring Gulls in Rhode Island and have a mean clutch size of 2.83.

Black-hacked Gull eggs hatched first in all Herring Gull colonies. Hatching success
on all islands was high (95%) compared to that reported by Harris (76%. Ibis 106:
432-456, 1964) and Erwin (44%, ibid). The high hatching success on Clam Island
may he due to the lack of mammalian predators and the nesting synchrony in the area.
Herring Gulls and Black-hacked Gulls established territories at about the same time
in these epicenters, thus eliminating behavior in defense of their nests which would have
a disruptive effect. Erwin (ibid) attributed the low success to excessive territory defense
on the part of the Black-hacked Gulls because of their synchrony with the Herring Gulls.
Within 10 days of hatching in New Jersey, the chicks entered the nearby channels and
hays (5-20 m) when disturbed by humans. Generally the brood remained together with
one or both parents flying overhead. Parents did not mob the intruder hut left with
the chicks. Chicks usually remained near land hut sometimes swam 100 m away from
the islands. After the intruder left, the parents flew hack to the nest, landed, and
gave repeated calls. The chicks, visually isolated from their parents and nest, returned to
the nest within a half hour. Herring Gull chicks did not enter the water until consider-

GENERAL NOTES

305

ably older (over 25 days), and their parents moljbed human intruders. Perhaps one
advantage of Black-hacked Gulls nesting with Herring Gulls is the protection the at-
tacking Herring Gulls provide.

Although Black-hacked Gulls are well known predators on the eggs and young of
other gulls and terns (Hatch. Auk 87:244-254, 1970), I did not observe any instance of
Black-hacked Gull predation on either eggs or chicks in over 560 h of observation from
a hide where I could see 40 Herring Gull nests and 2 Black-hacked Gull nests. Similarly
I never found the remains of eggs or young near the 8 Black-hacked nests regularly
checked on Clam Island. Even though over 500 Herring Gulls were handed in the
vicinity. Black-hacked Gulls never regurgitated any hands near their nests. The mean
clutch size of Herring Gulls was similar in areas with and without nesting Black-hacked
Gulls on Clam Island.

Black-hacked Gull adults generally ignored the Herring Gulls nesting nearby, although
the Herring Gulls appeared to move away from approaching Black-hacked Gulls. Black-
backed adults actively chased Herring Gulls only when Black-hacked Gull eggs were
hatching. When disturbed by a human, gulls circle overhead. During these distur-
bances, the Black-hacked Gulls chased and pecked at any other gull flying over their
nests and eggs.

Nine times I observed Herring Gulls initiate and successfully chase Black-hacked Gulls
when their chicks (as old as 32 days) approached adult Black-hacked Gulls. I saw only
1 fight when a Herring Gull chick walked to within 1 m of a Black-backed nest and was
chased by the resident adult. A fight ensued hetw'een the parents which involved peck-
ing, wing flapping, wing tugging, and feather pulling. The Black-hacked Gull subse-
quently flew and the Herring Gull chased its chick hack the 5 m to its own nest.

Black-hacked Gulls had high fledging success on the islands examined, as all pairs
fledged 2 to 3 young (X = 2.42). \oung were considered fledged once they reached
1300 g or w'ere able to fly. Young from Clam and Caravel islands were weighed periodi-
cally and their weights were similar to those reported by Harris ) ibid ) . It is difficult to
obtain fledging rates for this species since they are often eliminated from mixed species
breeding colonies by investigators interested in low' predation rates (Harris, ibid). The
high fledging rate in Ocean County, New Jersey may he the result of experienced breed-
ers, high food supply, low' predation rates, little conspecific competition for nest sites, or
a combination of these factors. Mammalian predators are lacking from these low tide
swept islands, and Herring Gulls do not yet appear to he as active predators on eggs
as usual.

I would like to thank W. Vesterman for critically reading the manuscript and F. Lesser
for valuable discussions and logistical support. — Joanna Burger, Dept, of Biology, Liv-
ingston College, Rutgers JJniv., New Brunswick, NJ 08903.

Wilson Bull., 90(2), 1978, pp. 306-308

PRESIDENT’S PAGE

(The following address was presented by Douglas James, President of the Wilson
Ornithological Society, first at a recent conference on the amateur in ornithology held
at the Cornell Laboratory of Ornithology, and later at the annual meeting of the Society
this past May in West Virginia.)

Among the diverse facets of science and natural history, the study of birds has been
particularly captivating to the lay person, tlie so-called amateur. Because of this rather
unique and continuing interest, the science of ornithology has benefited through favor-
able public support. In return, the interested public has enjoyed the good feeling as-
sociated with becoming involved. The ornithologist seldom has fostered a schism of
alienism that often pervades other scientific disciplines. Some would argue that this
accessibility of ornithology to the general public actually has retarded progress in
ornithology due to the inertia against changing concepts frequently encountered in the
views of the non-scientist. This, of course, is a debatable point. A consideration that is
not controversial, however, is the acknowledged large contribution lay members of
ornithological societies make through dues payments in supporting the ornithological
journals. Thereby the amateur greatly expands the opportunity for publishing the
fruits of scientific labors produced by the practicing ornithologist. This has to count
as an outstanding contribution of the amateur to the success of ornithology. Admittedly,
this large lay audience may exert a subtle negative influence on the quality of the journal
contents, but this factor certainly must be outweighed by the implemented great increase
in number of journal pages. Because of this kind of amateur support, there are 5
ornithological journals in this country that receive national and international circulation.
This is an enviable position not enjoyed by any of the sister sciences in vertebrate
l)iology. On inspection of membership rolls of the societies supporting these ornithologi-
cal journals, it is strikingly ol)vious that the proportion of amateurs as members is much
higher than the proportion of amateurs that publish in the same journals. The profes-
sional ornithologist is overrepresented in the publications and thus benefits greatly
from support from amateurs.

The Wilson Ornithological Society is one of these journal societies. An appraisal of
its activities clearly indicates that its main focus pertains to maintaining a quality |
journal in ornithology. The Wilson Bulletin. The only coordinated group activity of the
Society is the annual meeting, which is attended by merely 10% of the members. This !
is comparable to meeting attendance ratios of the other major ornithological societies. ,
Therefore, the Wilson Ornithological Society, like the others, reaches its membership, j
stimulates their interest, and attracts their support primarily through the publication |
of a journal devoted to papers in original research in ornithology. I cannot emphasize {
enough the importance to ornithology of the amateur’s role and interest in supporting !
this enterprise. t

Recognizing this present day situation, it is interesting to note that The Wilson I
Ornithological Society was indeed founded l)y amateurs, and by perhaps the most |
uninitiated cadre of amateurs, depending on one’s view. In fact, the charter members in |
1888 were a group of pre-college boys. The details have been stated l)efore, so I need j
not dwell on them here, except to say that many of these 36 youths previously had become i
acquainted in an organization called the Young Ornithologists’ Association. Most
originally had met through correspondence in response to egg exchange advertisements in i
The Oologist, but they were interested in more than just data gathering, and lacking a

306

PRESIDENT’S PAGE

307

deep knowledge of anatomy, physiology, and taxonomy, they wanted to conduct mean-
ingful studies in field ornithology. Even in those early years the membership had a
coast to coast scattered distribution united only by a publication that eventually became
The Wilson Bulletin.

All this emphasizes the early tradition of the amateur in ornithology in the Wilson
Society. And I purposely say “the amateur in ornithology” which is vastly different
from what would have been incorrectly implied by the phrase “amateur ornithologist.”
Some of these lads eventually became professional biologists. George Hall, who has
reviewed all the past issues of the Bulletin, informs me that this prevalence of the
amateur characterized the Society until the time of World War II. After that, the
proliferation of professional ornithologists was conspicuous and it continues to be a
significant trend in the Society.

So far I have described how the Wilson Society was at first a wholly amateur enter-
prise, becoming transformed more recently into a noteworthy support of ornithology
through amateur participation in Society membership. This latter, of course, is a passive
relationship. The Wilson Ornithological Society also provides an active support of the
role of the amateur in ornithology. I think more properly stated, it provides the amateur
with the opportunity to become significantly involved in and make contributions to
ornithology. The Society directly encourages ornithological research by amateurs through
the annual Margaret Morse Nice Awards, which are grants in support of research given to
deserving applicants who are not professional ornithologists. The annual Edwards Prizes
for the most significant papers in The Wilson Bulletin reward amateurs when they
happen to be the prize winning recipients. And, in fact, the Bulletin is an avenue for
publication of research projects conducted by amateurs, whether members or not. Many
amateurs have taken advantage of the open forums at annual meetings of the Society to
present the results of their research on the program for scientific papers. Also, amateurs
frequently are officers of the Society, and thereby play an especially significant role in
ornithology. Finally, the Society offers a particularly important opportunity to amateurs,
and all members, in maintaining the Van Tyne Library at the University of Michigan.
Holdings in avian biology in this library collection are available on loan to any society
member; this provides access to items that are not available in most local libraries
used by amateurs.

Everything about the encouragement of amateurs by the Wilson Society, and the
involvement of amateurs in the Society, speaks to a personal rather than a coordinated
group participatory relationship. There are no programs comparable to the Christmas
Bird Counts, Breeding Bird Survey, bird handing efforts, or nest record schemes,
specifically originated and coordinated l)y the Society. 1 do not view this as a mindless
deficiency on the part of the Society. The amateurs in the Society who have made really
significant ornithological contril)utions have described to me how their inspiration de-
rived from personal contact with scientists in the Society, made possible of course by
the existence of the Society, hut not arising from any special group program sponsored
by the Society.

Therefore, Kenneth Parkes is correct in reminding me that the Wilson Society
has been very instrumental in encouraging a»nateurs on an individual basis, but has not
been especially concerned with developing group participation projects that mobilize
amateur efforts focusing on s])ecific ornithological problems. Nevertheless, elsewhere
there certainly have been significant group efforts of this kind, more should be developed,
and The Wilson Ornithological Society would be pleased to offer its facilities to assist
in any way it can.

THE WILSON BULLETIN • Vol. 90, No. 2, June 1978

In view of the successful experience of tlie Wilson Society with regard to the participa-
tion of the amateur, I would make a recommendation concerning group projects in
ornithology. I think that the reason the Wilson Society survived and flourished under
the blend of amateur and professional expertise is that both groups realized a common
goal of accomplishment. This means that amateurs like professionals were encouraged
not only to gather pertinent data relative to their studies, but also could follow through
in data analysis, and could develop the conclusions pertaining thereto. If more group
projects are forthcoming that involve the mobilization of amateurs to collect ornithological
data relating to important avian research problems, then the same amateurs, to really
become involved and motivated, should have the opportunity in these projects to partici-
pate in the data analysis, data interpretation, and in forming appropriate conclusions.
This to me is the real lesson arising from the past experience of The Wilson Ornithological
Society and is the overriding dominant principle that spells success to the cooperation
between the amateur and professional in producing significant advances in the science
of ornithology.

ORNITHOLOGICAL NEWS

NOTICE TO CONTRIBUTORS

Effective immediately, new manuscripts to be considered for publication
in the Wilson Bulletin should be sent to editor-elect. Dr. Jon Barlow, Depart-
ment of Ornithology, Royal Ontario Museum, 100 Queen’s Park, Toronto, On-
tario, Canada MSS 2C6. Correspondence concerning volume 90 of the Wilson
Bulletin should be sent to the present editor.

HAWK MOUNTAIN RESEARCH AWARD

The winner of the first annual Hawk Mountain Research Award was James C. Bednarz
of Iowa State University for his studies of the “Status and habitat utilization of the
Red-shouldered Hawk in Iowa.”

The Board of Directors of Hawk Mountain Sanctuary Association announces its sec-
ond annual award of $250 for support of raptor research. The Hawk Mountain Research
Award is granted annually to a student engaged in research on raptors ( Falconiformes) .

To apply, students should submit a description of their research program, a curriculum
vitae, and two letters of recommendation by 31 October 1978 to: Mr. Alex Nagy, Hawk
Mountain Sanctuary Association, Route 2, Kempton, Pennsylvania 19529.

A final decision will be made by the Board of Directors in February 1979.

Only students enrolled in a degree granting institution are eligible. Both under-
graduate and graduate students are invited to apply. Projects will be judged com-
petitively on tbe basis of their potential contribution to improved understanding of
raptor biology and their ultimate relevance to conservation of North American hawk
populations.

Wilson Bull, 90(2), 1978, pp. 309-321

MANAGEMENT OF NATIONAL WILDLIFE REFUGES IN THE
UNITED STATES: ITS IMPACTS ON BIRDS

National Wildlife Refuges administered by the IJ.S. Fish and Wildlife Service, De-
partment of Interior, occur in 49 of the 50 states and encompass more than 13,678,860 ha.
While much of the present refuge system was acquired for management of migratory
waterfowl, refuges have been acquired for preservation of '‘endangered” species ranging
from Whooping Cranes iGriis americana) to the Dusky Seaside Sparrow ( Ammospiza
nigrescens) . Refuges have been acquired through withdrawal from the public domain,
donations, outright purchase, leases, easements, and acceptance of lands administered by
other agencies. Consequently due to the diversity of habitats and species, origin, location,
etc.. National Wildlife Refuges have different values to different interest groups. Despite
having wildlife-oriented missions, some refuges have been managed for grazing, recreation
such as boating, lumber products, commercial crops, etc. with frequent adverse effects on
achieving desired wildlife objectives. Multiple and single uses of refuges contrary to
initial objectives when refuges were acquired have resulted in internal and public
criticism. These problems coupled with inadequate funding and staffing have led to
outside review of the overall system (Leopold 1968). More recently intense dissatisfac-
tion with the LI.S. Fish and Wildlife Service’s (USFWS) “management by objectives”
approach to budgeting (resulting in no program specifically for refuges, continued in-
adequate funding of the refuge system, and a host of other alleged problems) has led
to much internal and public commentary on the desired future of the refuge system.
These problems led to preparation of draft and final environmental statements concerning
operation of the National Wildlife Refuge System (USFWS 1976b). During 1976
and early 1977 The Conservation Committee of The Wilson Ornithological Society
solicited comments on and reviewed major practices on National Wildlife Refuges. The
complexity of the refuge system, funding restraints, inherent operational problems,
legislative authority, etc. were such as to overwhelm the Committee. Consequently it
was decided to identify major practices on refuges affecting birds that could conceivably
he altered to enhance avian habitats and populations. The report relies heavily on the
waterfowl literature as data concerning raptors, colonial waterbirds, and other non-game
birds on refuges were generally not available,

HISTORICAL PERSPECTIVE

President Theodore Roosevelt, by executive order on 14 Alarch 1903, set aside Pelican
Island as the first federal bird refuge. By the end of his first term in 1904. Roosevelt
had created 51 wildlife refuges in 17 states and 3 territories. The Weeks-McLean Bill,
attached as a rider to the Agricultural Appropriations Bill and signed unknowingly by-
outgoing President William Howard Taft, gave the federal government authority over
migratory birds in March 1913. The intent of the Weeks-McLean Law. considered an
unconstitutional invasion of state’s rights, was given added authority by the Migratory-
Bird Treaty between the United States and Great Britain (for Canada) in 1916. Then,
in 1918 Congress passed and President Woodrow Wilson signed the Migratory Bird
Treaty Act.

The various treaties and laws were regulatory and, although migratory species
responded with increased numbers for a time, it became clear that long-range problems
and solutions were in protection of habitat. Refuges established by executive order were
too few and scattered to insure the future of migratory species. The first attempt to

309

310

THE WILSON lUJLLETIN • VoL 90, No. 2, June 1978

launch a program came in 1921 with l)ills that would estal)lish a refuge system, a
Migratory Bird Refuge Commission, and a one-dollar federal hunting stamp. Labeled
as a “duck slaughter” bill, it failed 4 times in Congress. Finally, in 1929, a hill passed,
hut only after stripping it of any shooting ground provisions and the federal hunting
stamp. It was to he funded with Congressional appropriations.

Concern for migratory species, especially waterfowl, increased as their numbers de-
clined with the drought of the 1930’s. Congress failed to appropriate funds for the
refuge system authorized in 1929. As a result of increasing concern, the federal hunting
stamp proposal was revised and finally passed in 1934. With a source of revenue and
the leadership of Jay “Ding” Darling, the National Wildlife Refuge System advanced
from a few scattered units to the system of 367 refuges that we have today.

Legislative authority, executive orders, and international treaties have given the federal
government responsibilities for all species occurring on refuges. These range from elk
{Cervus canadensis) on the National Elk Refuge in Wyoming, to the endangered Whoop-
ing Crane on the Aransas National Wildlife Refuge (NWR) in Texas. But, for ail these
responsibilities, the refuge system’s major focus and objective has been the preservation
and management of waterfowl. At the fiscal year 1974 level, 276 (75%) of all refuges
were managed specifically for Avaterfowl production, migration, or wintering. Wildlife
and Game Ranges and Big Game Refuges have principally been established by with-
drawals from the public lands. Over 9.5 million hectares in 24 units are in this category.
On 68 refuges, encompassing over 1.6 million hectares, management must he directed
tow^ard certain species of colonial nesting birds.

Maintenance of the National Wildlife Refuge System has not kept pace Avith the early
interest shown in its establishment. During the system’s expansion more than 11,000
people Avere engaged in developing critical Avildlife habitat. But, this support was not
so much for the refuges and their wildlife, as it was for the Civilian Conservation Corps
and the jobs it created during the Depression. Recently, USFWS Director L. A. Green-
wait testified, “The National Wildlife Refuge System, as with most activities of our
Service, has been underfunded for some time. The consequences are evident in facilities
which are inadequate and poorly maintained. Too few people are available to do a
proper job of refuge management.” Much needed funding and personnel for the refuge
system have been diverted to new responsibilities- — ^energy research and development,
Avilderness studies. Youth Conservation Corps, endangered species, and marine mammals.
These activities have not been funded on their OAvn merit, hut at the expense of the
National Wildlife Refuge System. During the 1975 fiscal year, refuge field operations
Avere funded at about $20 million, $7.1 million less than the 1970 funding level. The
USFWS has estimated that, to fully develop the entire system to provide optimum
wildlife and public benefits, $170 million would be needed. To maintain that level of
operation, an additional $34 million and 2000 man-years of labor Avould he needed
annually. The Carter administration has recommended a 30% increase in the level of
funding for the USFWS. It proposes additional personnel ceilings under the Bicentennial
Land Heritage Program.

MAJOR REFUGE MANAGEMENT ACTIVITIES THAT AFFECT BIRDS
Grazing

According to the Final Environmental Statement on the Operations of the National
Wildlife Refuge System (USFWS 1976b), in Fiscal Year 1974 (1 July 1974-30 June
1975) 740 grazing permits were issued to private citizens for approximately 526,110 ha

CONSERVATION COMMITTEE REPORT

311

of refuge land. These 740 permittees used 354,589 Animal Unit Months (AUM’s).
Grazing occurred on 103 refuges in 36 states, using 4% of the total area on National
Wildlife Refuges. Areas grazed in addition to rangeland included native wet meadows
and riparian sites which are used extensively by nesting waterfowl and other birds.

The number of hectares grazed in each fly way were: Atlantic — 5,947, Mississippi —
10,360. Central — 158.321, and the Pacific — 331,887. In addition, about 6475 ha were
grazed in Alaska. Forty-five percent of the 1,157,235 ha of rangeland in the system were
used by cattle. Nearly 70% of the total grazed area occurred in 3 states: Montana (8
refuges), Nevada 15 refuges), and Oregon 14 refuges).

Although a logical assumption would he that the high grazing use in the Pacific
Flyway mostly occurs on big game refuges, such as Hart Mountain NWR, Oregon,
Sheldon NWR, Nevada, and National Bison Range, Montana, such is not the case.
Malheur NWR, Oregon, a waterfowl and waterbird production area, had 98,502 AUM’s
in 1974-75. or 27.8% of the national refuge AUM total. Hart Mountain NWR had only
11,000 AUM’s or 3.1% of the national total. Sheldon NWR, w^hich is considered over-
grazed, had 24,000 AUM’s 16.8%) while the National Bison Range had no cattle grazing.

The present USFWS policy is that grazing and haying programs he used to manipulate
vegetation to maintain or increase wildlife productivity and species diversity over a
sustained period of years at minimal cost to the government, and that grassland habitat
should he maintained for the primary benefit of wildlife populations. Grazing and
haying activities may be permitted to enhance, support, and contribute to established
wildlife management objectives, but must not conflict with those objectives lUSFWS
1976b). Unfortunately, when grazing is allowed, the USFWS frequently loses control
of local situations due to intense political pressure at all levels of administration.

At least 55 waterfowl studies have shown that grazing is detrimental to waterfowl
production. Only one study reported higher success on moderately grazed areas than on
idle areas 1 Burgess et al. 1965). Anderson H957) reported that 42.2% of the 116 nests
on idle land in California hatched, while none of 7 on grazed land hatched. Glover
0956) found 24.4% nesting success on idle land and lightly grazed areas in Iowa,
compared with 10.5% success on moderately and heavily grazed areas. One study had
nest losses of 80% in light cover, compared with 29% in dense cover fSchranck 1972).
Weller et al. 0958) reported that the effect of cattle grazing on vegetation in Utah was
as serious as the lack of water. On Malheur NWR, Oregon. Greater Sandhill Crane
iGrus canadensis tahida) nesting success in 1976 was 54,6% in mowed-grazed. 63.6%
in mowed-ungrazed, and 84.2% in unmowed-ungrazed areas fC. D. Littlefield, unpubl.
data) ,

Grazing was reduced sharplv in some regions of the United States after a memorandum
was released 22 December 1972 from the Director. USFW^S. It stated “Recent research
at the Northern Prairie Wildlife Research Station at .Tamestown. North Dakota, indicates
that having and grazins are incompatible with upland nesting duck and ground nesting
bird objectives. . . . While the recommendations apply primarily to the north central
region of the United States and the southern Prairie Provinces of Canada, application
of these practices on refuges in other geographic areas have demonstrated similar
favorable response by w'aterfowl and other ground nesting birds.”

In North Dakota. AU^^s were reduced after this memo, but by 1975 had increased,
with additional increases planned in the future. At J. Clark Salyer NWR. North Dakota,
three-quarters to 1 AUM per acre C4 ha) was used in 1976 Dotal 2600 AUM’s), but
present plans are to increase the use to 2 AUM’s per acre. Other examples in North
Dakota include Arrowwood NWR with 435 AUM’s in 1971; 1109 in 1975; and 16.50

312

THE WILSON BULLETIN • VoL 90, No. 2, June 1978

projected for 1979. Upper Souris NWR had 2348 AUM’s in 1971; 2348 in 1975; and
5634 are projected for 1979.

The most serious grazing i)rol)lems on National Wildlife Refuges appear to occur
in Oregon and Nevada because of local political pressure with over- or untimely grazing
being typical of most western refuges. Examples given are hut a small sample of the
problems associated with this management “tool.”

In 1948 on Malheur NWR, waterfowl production was 150,000 ducks, hut in 1974 only
21,300 were produced. In 1948 AUM’s were 74,385, increasing to 101,726 by 1951. In
1961 and 1971 AUM’s were 122,404 and 123,807 respectively. As AUM’s increased duck
production decreased. From 1962 to 1972 the average number of ducks produced an-
nually was 29,600. Mallards i Anas platyrhynchos) , which are dependent on residual
vegetation from the previous year for nesting cover, declined from 50,000 produced in
1949 to 2,120 in 1974. Some changes in grazing practices are presently occurring at
Malheur. By 1975-76 AUM’s had been reduced to 88,221. After considerable pressure
from environmental groups in 1976, AUM’s were reduced to 65,828. In addition, 1712 ha
were mowed for hay. By 1977, 8782 ha were in “non-use,” compared with 263 ha in 1962.
At Malheur NWR the grazing program requires over 338 km of internal fences. In 1976,
to protect river hanks and dikes from severe cattle trampling, several km of additional
3-wire fences were placed between heavily grazed areas and canal and river banks.
These new fences have resulted in many Sandhill Crane (Grus canadensis) pairs having
their traditional territories bisected with barbed wire. Obviously, grazing in this situa-
tion does not enhance refuge objectives but instead creates conflicts with adverse effects
on bird populations.

Wildlife collisions with fences are common. Mule deer [Odocoileus hemionus) ,
pronghorn ( Antilocapra arnericana ) , and numerous birds have been killed flying into
or becoming entangled in fences. Greater Sandhill Cranes have been killed at Malheur
and Grays Lake NWR, Idaho. Flightless young Whooping Cranes became entangled in
barbed wire fences on a number of occasions at Grays Lake NWR, in 1975. One young
Whooping Crane died in 1976 near Monte Vista NWR, Colorado, after colliding with a
fence. At Red Rock Lakes NWR, Montana, several moose iAlces alces) calves have
died from injuries sustained after becoming entangled in fences. At times, simple
modification of fences can he beneficial. While over 20 km of interior fences have been
removed at Grays Lake NWR, virtually all of the remaining fences have been modified
from 4-5 to 3 wires. This has greatly improved movements of young cranes. Of im-
portance is the obvious fact that fences have little value for wildlife. They are expensive
and are placed on refuges primarily to enhance livestock grazing.

On many refuges, power lines transect nesting areas or bisect principal flight paths.
Some of these power lines bring electricity to pumps that supply water for cattle. Power
lines are a major mortality factor for swans, cranes, eagles, and other large birds. At
Bosque del Apache NWR, New Mexico, aircraft markers have been placed on lines
and some lines have been buried. Mortality has been dramatically reduced. On other
refuges either the pumps should he removed or power line markers should be placed
at strategic locations in wildlife use areas. Preferably the lines should he removed or
buried.

At Stillwater NWR, Nevada, the USFW’S operates the refuge with a cooperative
agrc(‘ment between the Bureau of Reclamation. Nevada Fish and Game Department, and
the Truckee-Carson Irrigation District. Much of the 90.653 ha refuge is unfenced and
cattle move onto the area freely. On the portion that is fenced, grazing is permitted for 11
months annually; total refuge AUM’s is 15.000. Refuge areas are leased from the
Bureau of Reclamation by the local irrigation district. Refuge personnel collect AUM

CONSERVATION COMMITTEE REPORT

313

fees, issue permits, and count livestock on and off the refuge. In return, all fees collected
are given to the irrigation district. Most grazing is determined by the district and
attempts to reduce AUM’s have failed ( one permittee is a member of the irrigation dis-
trict board ) .

Napier (19741 in his recommendation for Stillwater Marsh, stated “Grazing was used
as a tool for opening up shorelines overgrown with dense stands of cattail and bulrush
in Stillwater Marsh’s early history. The result was increased duck production. Now,
poor water conditions have resulted in a downward trend or elimination of emergents
on some impoundments. Management is now aimed at encouraging emergent aciuatic
growth. Livestock grazing in the marsh is detrimental in this respect, for cattle heavily
graze the emergent vegetation.” Because the Bureau of Reclamation, a sister agency of
the USEWS in the Department of Interior, owns the land, methods should be investi-
gated to solve not only the grazing problems, hut also the water deficiencies that pres-
ently exist on the refuge. Stillwater NWR is unique in that it provides marshland habitat
in an area that has few wetlands.

Summer grazing continues, although at a greatly reduced level from 1975 and 1976,
on Grays Lake NWR, Idaho, even though the endangered Whooping Crane is presently
being introduced by transplanting their eggs into Greater Sandhill Crane nests. In
1975 two young Whooping Cranes disappeared within 2-3 days after large numbers of
cattle were introduced into areas occupied by these chicks.

Improvements have been made on some refuges. Hart Mountain NWR, Oregon,
began reducing AUM’s in 1969 and in 1976 (11,000 AUM’s) the number of pronghorn
antelope young per 100 does was 59. On Sheldon NWR, Nevada C24,000 AUM’s), im-
mediately south of Hart Mountain NT^^R, the young-adult ratio was only 22/100 (E.
McLaury, pers. comm.). Present plans are to reduce the number of AUM’s at Sheldon.

At Bosque del Apache NWR, New Mexico, all grazing has been terminated. All internal
fences have been removed and many pastures which had been “improved” for cattle have
been converted to wildlife food crops or man-made marshes. Wintering Snow Geese
{Chen hyperhorea) have increased from a few hundred to over 21.000 in the past 10
years and Greater Sandhill Cranes have increased from 3200 to over 12,000.

At Ruby Lake NWR, Nevada, grazing occurs from 15 April through 1 January. Present
plans are to reduce the 5200 AUM’s by one-half. Wildlife changes that occur in the
deferred and hayed-only areas will be monitored and compared with those in areas that
continue to he grazed.

At Red Rock Lakes NWR. Montana, the management announced to local stockmen
that there would he a 10% reduction annually in AUM’s over a 5 year period. Sur-
prisingly, little opposition was encountered and the program is in its third year, with
AUM’s now 30% fewer than the original 13,144.

Prescribed burning has been used in grassland management to maintain desired suc-
cessional stages. To avoid the cost of fencing, issuance of permits, soil erosion, over-
fertilization f affecting water quality), and other aspects of grazing programs, burning
could he used to accomplish the same objectives. This would also prevent the refuge
system from becoming more involved with and influenced by local stockmen, grazing as-
sociations, and political pressure iVoight 1976).

Having:

In 1974-75, 16,714 ha were mowed for hay hut ungrazed by 589 permittees on National
Wildlife Refuges. In comparison with grazing, haying generally creates only minor
conflicts with wildlife management. The 3 major grazing states had minimal hay acre-

314

THE WILSON lUILLETIN • VoL 90, No. 2, June 1978

ages in 1974-75. The Central Flyway was most important, with refuges in North Dakota
(4521 ha) and Nebraska (4185 ha) I)eing leaders in this practice (USFWS 1976b).

In some situations haying can l)e beneficial. Native grasslands that receive flood
water in late spring can he mowed to discourage early nesting species. In areas with
limited water supplies, channels can he mowed to allow for rapid water movement.
Data from Malheur NWR, Oregon, have shown that Greater Sandhill Cranes, Canada
Geese (Branta canadensis) and some species of ducks, feed and loaf in mow'ed areas, but
prefer to nest in unmowed areas.

The major conflict with mowing is the time of year when it begins. Interviews with
mower operators on private land in southeast Oregon in 1976 indicated high mortality of
young birds from 1 to 15 July. Two operators estimated they had killed between 400
and 600 birds during this 2-week period. Most of these were shorebirds, but numerous
waterfowl nests, young ducks, and crane chicks were also reported destroyed. One
operator stated that he had killed 2 pronghorn antelope young in 1975. On Malheur
NWR, 4.2% of the Mallards hatch after 16 July. Other species and hatching percentages
are Gadwall (Anas strepera) 14.5%, Cinnamon Teal (Anas cyanoptera) 15.0%, and
Redhead ( Aythya americana) 13.4%. Many newly hatched broods are seen after 1
August. Younger Greater Sandhill Cranes suffer high mortality from mowing, especially
early in the season. Young cranes lie down and hide in vegetation when approached and
remain hidden until hit by the mower. Recently at Malheur NWR, haying has been de-
layed until 10 August to allow cranes time to fledge. In some areas on the refuge where
flightless young are known to occur, mowing has been delayed even longer.

Many refuges begin mowing activities in July, with some possibly as early as late
June. Until recently Medicine Lake NWR, IMontana, initiated mowing on 15 June.
Because of political pressure from one permittee, it was about 8 years before mowing was
terminated after it was publicly announced that it was to be stopped within 2 years.

Data collected at Malheur NWR is potentially applicable to other waterfowl produc-
tion areas. Refuges that allow mowing before 1 August are contributing to substantial
losses of wildlife. Biological data on the effects on wildlife should he collected on
refuges that have early mowing programs. To alleviate losses, haying should be delayed
until 15 August. It is important to note that virtually no data are available regarding
the impact of haying (or grazing) on other ground nesting birds.

Farming ;

Farming for production of cereal grains for waterfowl use has long been a major
endeavor on many refuges. Other crops (including oranges!) are sometimes grown. In
1974 at least 131 refuges farmed about 65.966 ha (USFWS 1976b). Primary reasons
for farming on refuges relate to providing supplemental foods for waterfowl during
migration and wintering periods and for preventing crop damage outside refuge
boundaries. The latter has not lieen overly successful when the large concentrations of
waterfowl, especially geese, cranes, and ducks on some refuges are considered. Farming
practices on refuges have been successful in concentrating birds, frequently too much |

so as witnessed by problems Avith shortstopping birds before traditional wintering areas are j
reached, crop damage problems adjacent to refuges, hunter firing lines leading to such j
prol)lems as lead-poisoning dieoffs and slob behavior of hunters, and outbreaks of density
dependent diseases such as foAvl cholera and duck viral enteritis. Possible negative
side effects of farming on National Wildlife Refuges may result from crop associated
use of pesticides and herbicides.

In the near past many refuges were evaluated on number of days of use they provided ||

CONSERVATION COMMITTEE REPORT

315

for waterfowl. Consequently the pressure was to show yearly increases in numbers of
waterfowl using areas over longer time intervals. It is now recognized by some refuge
managers and administrators that these goals were not beneficial to the waterfowl re-
source. Consequently, amount of land farmed on some refuges is decreasing with diversi-
fication from cereal grains becoming apparent. Goals of refuges should be reevaluated.
It is probable that having the bulk of individual populations of birds on one refuge
for long periods, such as is common with geese, is not healthful for the birds or beneficial
to the overall management of the resource. Diversified and well dispersed refuges,
especially in migration and wintering areas, are most desirable.

Timber management

Management of timber for the luml)er and pulp industries on National Wildlife Refuge
lands occurred on 21 refuges, primarily in the southeast and northeast in 1974. These
21 refuges reported a timber harvest from 12,141 ha (USFWS 1976h). While this may
be a small portion of the overall refuge system, impacts on some refuges are extensive.
As an example of the magnitude of these activities, the annual operating budget for
Noxubee NWR, Mississippi, has in recent years been about $110,000, yet this 13,760 ha
refuge has sold up to $250,000 worth of timber per year. The income goes into the
Federal Treasury and does not come hack to the refuge system. Timber management
and some economic gain from the forests on National Wildlife Refuges is not inherently
had. The extent and type of management may he. For example, the USFWS slogan used
to characterize timber management on southern National Wildlife Refuges is “all-age
management in even-age units.” This is a euphemism for clear-cutting. A booklet de-
scribing this management system on Noxubee NWR states: “The highly productive

alluvial soils (growing mostly hardwood) are managed under a long rotation (120 years)
and a frequent cutting cycle (15 years). The rotation age for upland areas (including
both pine and hardwood) is 80 years, and a cutting cycle is 10 years.” Thus, “all-age
management” allows some hardwoods to grow to the age of 120 and some pines to grow
to the age of 80. If such a plan was truly for “all-age” management, the rotation cycle
should he based on the natural potential longevity of the trees involved. Sizes of clearcuts
are stated in USFWS brochures to he limited to 12 ha though refuge foresters admit that
some cuts approach 20 ha. Aside from rotation ages and sizes of cuts, there is enough
controversy over the ecological effects of clearcutting (decreased diversity, etc.) that this
practice seems inappropriate for management of a National Wildlife Refuge. In short,
the forests of National Wildlife Refuges in the southeast are being managed to maximize
economic return at the expense of those wildlife species such as Red-cockaded Wood-
peckers iPicoides borealis) which require more mature forests. An important com-
ponent of the southern forest ecosystems is being lost.

Predator control

Few data are availal)le on predator problems on National Wildlife Refuges. In 1972
predator control through use of toxicants was discontinued on most public lands. Un-
fortunately, few comparative data were collected before 1972, and on most refuges little
has been collected since 1972.

It is possible that manv refuges have only minor predation problems, hut some have
high predation rates. Data have been collected on Greater Sandhill Cranes on Malheur
NWR, Oregon, since 196)6. After predator control through poisoning was terminated in
1972, production was greatly reduced fro?n 1973 through 1975. From 236 pairs that nest
on the refuge, only 2 young fledged in both 1973 and 1974. In 1975 only 17 fledged.

316

THE WILSON BULLETIN • Vul. 90, No. 2, June 1978

Common Ravens {Corvus corax) and raccoons {Procyon lotor) were the major egg
consumers, while coyotes {Canis latrans) took eggs and young. In the winter of 1972-73
the hlack-tailed jackrahhit i Lepus calijornicus) population sharply declined and coyotes
moved onto the refuge to find alternate food sources. One pond where more than 300
young Canada Geese were captured and handed in 1972 produced only 28 in 1973. It was
not uncommon during mid-day in August to observe more than 45 coyotes along one 68
km road through the refuge. Coyotes normally cause only minor problems on Malheur
when jackrahhit numbers are high. However, Common Ravens find ideal conditions in
southeast Oregon. Numerous rimrocks provide nesting sites and the cattle industry
and nesting birds provide an abundance of food. One roosting site on Malheur Lake in
1976 was being used by more than 800 ravens.

Two noteworthy predation incidents were documented in 1976. At Crescent Lake NWR,
Nebraska, 2 of 5 Trumpeter Swans iOlor buccinator) that were to he released in 1977
were killed by raccoons or coyotes, and 26 Greater Sandhill Crane nests including 3 that
contained transplanted Whooping Crane eggs, were destroyed by coyotes at Grays Lake
NWR, Idaho. Predator problems have also been reported at Attwater Prairie Chicken
NWR, Texas.

Many refuges are artificially developed with numerous canals, artificial ponds, nesting
islands, water control structures, and other man-made elements to attract waterfowl and
other birds and encourage nesting. Such an artificial environment also attracts large
concentrations of predators, especially when predator control is being practiced on sur-
rounding private lands. With habitat manipulation, species that require dense nesting
cover are henefitted. But for species that nest in open situations and construct nests that
are exposed during periods of absence, dense vegetation is of little value, and nests are
especially vulnerable to avian predation. Many species of shorebirds and marsh birds
fall into this category. On refuges that support breeding populations of species with low
reproductive potential, predator management should be used to insure their continued
survival.

Severe losses on some refuges will continue if predator populations remain unchanged.

If nesting studies are not initiated to ascertain the impact of predation, present practices
of non-control will continue. Whether California Gulls iLarus calif ornicus) in Utah,
Black-hilled Magpies iPica pica) in Colorado, striped skunks (Mephitis mephitis) and j
red foxes iVulpes fulva) in North Dakota. Common Ravens in Oregon, etc., the impacts ]
of predators on federal wildlife refuges need to he examined. If predators are a serious |
detriment to the production of other wildlife, their populations should be properly
managed.

Recreational activities

Fishing. — In 1974 fishing was allowed on 171 refuges with fishing waters being stocked
on at least 18 refuges (USFWS 1976h). Generally fishing is a recreational use of refuge
wetlands that is compatible with the protection and management of birds. However,
excessive use of shallow vegetated areas of lakes and streams by wading and boating
fishermen can disturb feeding and nesting waterhirds. Many southern refuges, such as
Noxubee NWR, prohibit fishing during the winter months in order to provide sanctuary
for wintering waterfowl, though when nesting activities of resident species are beginning,
the lakes are opened to fishing again. Prime nesting areas on many refuges are closed
to fishing until about 1-15 July. Such dates are unrealistic on some refuges as nesting
continues after these dates. Timing of fishing closures ( if any ) varies from refuge to
refuge and no policy appears to have been formulated on this use of refuges above the

CONSERVATION COMMITTEE REPORT

317

local level. In northern areas, fishing should he delayed until about 1 August with some
presently open fishing waters being closed to protect late nesting species and their broods.
In general fishing regulations on refuges are appropriate but some refuges allow use of
trot lines upon which mergansers, loons, and diving ducks have inadvertantly been
snared. This is an unnecessary abuse of National Wildlife Refuges.

Boating. — Various sizes and types of boats have been used on National Wildlife
Refuges for many years in pursuit of refuge management goals and fishing. With the
advent of motors and more leisure time, various publics have demanded and received
access to National Wildlife Refuges for motor boating and water skiing. Presently 42
refuges permit high speed pleasure boating; mostly on areas where the USFWS has
secondary control (USFWS 1976b). Obvious and documented impacts of high speed
boating are shoreline degradation, disruption of nesting and feeding areas with loss
of production of young, and displacement of water birds. These problems, especially loss
of production of young, are especially pronounced at Ruby Lake NWR, Nevada, and
have resulted in the preparation of an Environmental Impact Assessment on the effects
of boating at this refuge (USFWS 1976a).

Ruby Lake NWR was established 2 July 1938 as a refuge and breeding ground for
migratory birds and other wildlife. Most of the 15,229 ha area was purchased, with the
remainder being withdrawn from the public domain. Boating was allowed for the pur-
pose of fishing starting in the mid-1940's, with water skiing being allowed starting about
1955. Sizes of boats and motors and numbers of water skiers increased until the late
1960’s when the USFWS moved to protect the waterfowl resource (principally nesting
Canvasback, Aythya valisineria, and Redhead ducks) by restricting power boating.
Since that time public and political pressures have prevented adecjuate restriction of
power boating necessary to protect the waterfowl resource with concomitant decreases
in production of over water nesting waterfowl. More recently commercial developments
by large corporations have resulted in a proliferation of sub-divisions for recreational
homes in the area near the refuge. Advertisements clearly indicate that Ruby Lake NWR
and associated water related activities on the refuge are important inducements attracting
people to purchase “ranchettes,” etc. With increasing political and public demands for
beating related activities on this refuge, it is (juite obvious that the original purpose
of the area has been lost. The future of this refuge is in dire straits and it may become
a recreation area if public apathy cannot be changed to prevent local abuse of a national
resource.

Boat related disturbances with no or little consideration of values of wetlands and
associated water birds have no place on National Wildlife Refuges. When threatened
or endangered species are impacted by such activities, closures of refuges to boats should
be mandatory.

Hunting. — Sport hunting of wildlife was permitted on portions of 184 National Wild-
life Refuges in 1974 (USFWS 1976b). Hunting was ])riniarily for migratory waterfowd
but also w^as allowed for resident game birds and big game species. Since hunters have
provided funds for much of the prime wildlife habitat purchased for refuges, it is logical
and rational that some level of hunting be allowed. Few refuges are completely open
to sport hunting and it would appear that state and federal regulations on season
length, hag limits, methods of taking, etc. are more than adequate to maintain avian
resources. Where endangered species are involved, such as Whooping Cranes and
Mexican Ducks ^ Anas diazi). it is difficult to see the rationale for sport hunting of look-
alike species. Hunting of look-alike species on those few refuges where these potential
problems exist should necessarily be reevaluated and probably discontinued.

318

THE WILSON 8ULLETIN • VoL 90, No. 2, June 1978

Other management problems

A recurring prol)lein on National Wildlife Refuges is the ease with which refuge lands
can he abused by other federal agencies. Some refuges have been used as practice
areas for low flying military aircraft, others as convenient and inexpensive routes for
highway and utility rights-of-way. The advent of NEPA hopefully will eliminate some
of this abuse, but problems still remain. For example, the USFWS had no olijection
to the channelization of tlie Yazoo River by the U.S. Army Corps of Engineers through
\azoo NWR, Mississippi.

A recent trend in refuge management has been to consolidate land holdings and to
adjust refuge l)oundaries to facilitiate management. This has the effect of making a nice
compact refuge rather than one with “fingers” extending into the surrounding non-refuge
lands. While we appreciate the management problems involved, such consolidation often
results in losses for wildlife. At one refuge the adjustments in land holdings were made
by trading prime forest land for agricultural lands. In addition to the loss of not-soon-
to-be-replaced forest habitat, the deal also resulted in a net loss in acreage to the refuge.
Apparently it is easier to trade lands than it is for a refuge to either sell or purchase lands.
Hence, based on market values, the refuge traded more acres of forest to obtain fewer
acres of crop land. We feel that the increased edge and linear distances on more dis-
persed refuges can often provide habitat for larger wildlife populations than could
compact refuges. An added benefit of such dispersed refuges is that they often provide
an ecological archipelago that will allow wildlife the opportunity to disperse to other
suitable habitats outside the refuge.

Some National Wildlife Refuges include areas of potential value as wilderness. Such
areas should be identified and protected. Personnel at one refuge indicated that such
an area occurred on their refuge, but that they were going to construct a road through
the middle cf it so that it would not qualify for wilderness status and so that current forest
management practices could be continued. Such actions are deplorable.

In addition to management or lack of management on National Wildlife Refuges
that affects birds, we feel compelled to point out a few refuge “management” practices
that adversely affect ornithologists and bird-watchers. Refuges tend to be generally j
understaffed as a result of inadequate funding. One reflection of this prol)lem is the
operating schedule for most refuges. Refuge offices typically open at about 08:00 and i
close about 16:30 Monday through Friday and are closed on weekends. This is fine '
for carrying out wildlife management activities, but many refuges also receive large
numbers of human visitors — most on weekends and after regular working hours. An
open office with descriptive brochures and bird checklists could win a lot of support
for the refuge system. Additionally, as some refuge managers see it their biggest prob-
lems are managing people. Perhaps these management problems would be fewer if
refuge public relations were improved by tailoring refuge office hours to accommodate
visitors and by providing informational materials.

Ornithologists seeking to conduct ecological research on National Wildlife Refuges
are faced with an unwarranted numl)er of bureaucratic prol)lems. Not only are state and j
federal bird banding permits required, but the researcher must also obtain a refuge permit j
and file an annual report of his activities on the refuge. If permits were simply obtained '
by visiting or writing to refuge headquarters, the requirement would not seem unreason- i
able, but often this is not the case. Permit requests are often channelled through ; '
regional USFWS offices, sometimes through Washington, D.C., before a permit is
granted — thus causing the researcher loss of valuable time. Collecting permits for refuges '
are particularly difficult to obtain and perhaps justly so, were it not for the fact that |

CONSERVATION COMMITTEE REPORT

319

hunters are often given freedom to hunt game birds on refuges with no more difficulty
than obtaining a state hunting license and a duck stamp. Another problem associated
with conducting research on National Wildlife Refuges is the USFWS practice of
frequently moving personnel from one refuge to another. As soon as a researcher has
established a good working relationship with one refuge manager, he is often con-
fronted with explaining his work and adjusting his research activities to conform to a
new manager’s interpretation of regulations. In all fairness, however, the legitimate
ornithological researcher has much to gain from working on National Wildlife Refuges.
Our experience has for the most part been that refuge personnel are eager to have
research conducted on refuges and that they are willing to provide logistic support
whenever possible.

SUMMARY

National Wildlife Refuges administered by the U.S. Fish and Wildlife Service, De-
partment of Interior are located in 49 of the 50 states and encompass more than 13,678,860
ha. While purchased or obtained for a variety of purposes including migratory birds
(primarily waterfowl) and endangered species. National Wildlife Refuges are vitally
important for maintenance of important habitats and overall conservation of many
species of birds. Problems associated with management of National Wildlife Refuges
include: (1) concentrating large numbers of birds which increases risk of catastrophic

losses due to disease and other mortality factors and the opportunity for damage to
items valued by man; (2) overgrazing by domestic livestock; (3) cropping for hay;
(4) water oriented activities such as boating; (5) creation of monocultures by selective
cropping or planting practices; (6) a lack of selective management of predators; (7)
failure to consider impacts of artificial structures such as fences, powerlines, signs, etc.;
and (8) inadequate manipulation of biological and mechanical tools useful for main-
taining and imj)roving habitats useful for Idrds. Major administrative problems include
failure to clearly identify and support objectives of individual refuges and woefully
inadequate funding for refuge staffing and maintenance.

RE C O M !M EN D A TI ON S

1. Creation of a National Wildlife Refuge Service equal to the U.S. Fish and Wildlife
Service in the Department of Interior to manage National Wildlife Refuges would
result in unnecessary bureaucracy, diversion of talent and funding, and would
fragment a cohesive national policy for protecting habitat for wild animals. Ad-
ministration and management of National Wildlife Refuges should continue as a
function of the U.S. Fish and Wildlife Service, Department of Interior. However, the
refuge system should be given full program status and administration should be
streamlined with an Assistant Director directly responsible for the refuge program.

2. Funding for staffing and maintenance of National Wildlife Refuges has been woe-
fully inadecpiate for many years. Adequate funding to maintain refuges should be
strongly sui)ported. Funding for enhancement of existing refuges is desperately
needed as is funding for expansion of the refuge system. Funding should be in-
creased for public relations and hiring of non-game biologists.

3. Objectives of each National Wildlife Refuge should be reevaluated with manage-
ment being directed towards obtaining desired objectives once they are defined.

4. Uncontrolled grazing by domestic livestock has been documented to adversely affect
nesting success and productivity of birds. Grazing of domestic livestock on National

320

THE WILSON BULLETIN • VoL 90, No. 2, June 1978

Wildlife Refuges should be carefully evaluated and in many instances reduced to
levels compatible with refuge objectives. Proper timing of grazing is critical and
all summer grazing in production areas should be eliminated. Winter grazing
should be allowed only for desirable habitat manipulation where controlled burning
is not feasible. Refuges should not be managed for domestic livestock production.

5. Controlled burning has value for manipulating habitats and it should be further
tested with effects documented. Where beneficial, controlled burning should be
used in refuge management.

6. Unnecessary structures such as fences, powerlines, etc. should be removed within
refuge boundaries where they have been documented to be hazardous to birds. All
necessary structures should be marked with aircraft warning markers or other devices
to prevent and reduce bird-object collisions.

7. Mowing of habitats for hay crops or other refuge objectives should be delayed
until 1-15 August in production areas important to birds. Dates of mowing after
1 August should depend on locality and local condition. Management of refuges
for commercial hay crops is not desirable.

8. Selective control of predators on refuges managed for birds should be implemented
in areas where limited nesting and brood cover occurs or where severe local condi-
tions exist. Management of production refuges should seek to prevent ecological
situations favorable to maintaining or encouraging unnatural concentrations of
predators.

9. Excessive or unnatural fall and winter concentrations of birds should be discouraged
through habitat manipulation on refuges; such concentrations invite catastrophic
losses and damage to private property.

10. Public recreation activities on National Wildlife Refuges should not be given
preference over stated objectives of the refuges. Examples of undesirable activities
when birds are nesting are boating, water skiing, and fishing. Non-human use
areas are an integral part of the refuge concept and all human recreation activities
should be meshed within the objectives of each refuge. Public visitation should be
encouraged on portions of refuges with adequate staffing and suitable open hours.

11. Diversity of habitats should be encouraged on National Wildlife Refuges and
practices that lead to large areas of monoculture should be discontinued. This is
especially a problem in forested areas.

12. Forest management on National Wildlife Refuges should take into account the
natural potential longevity of the tree species present and should provide for the
needs of species chaiacteristic of mature forest ecosystems.

13. Collection and compilation of data concerning the effects of management practices
on avian species, especially non-waterfowl, should be an integral part of refuge
management. Research into management procedures and other scientific endeavors
should be encouraged on National Wildlife Refuges with improvement of permit
procedures and requirements being immediately instigated.

14. Consideration should be given where feasible to include portions of some refuges
in the Wilderness System to further protect unusual and unique habitats.

ACKNOWLEDGMENTS

The volume of correspondence generated by interest in this topic was considerable.
We thank all who corresponded with us and hope that confidences were not compromised.
Numerous personnel of the U.S. Fish and Wildlife Service assisted us in loeating litera-
ture references and in providing access to unpublished refuge reports. These individuals

CONSERVATION COMMITTEE REPORT

321

are dedicated to maintaining and enhancing wildlife habitats despite administrative
harassment, snafus, and inadequate funding. We are indebted to them and are sincerely
appreciative of their efforts. Special thanks go to E. M. Brigham, III, R. Drewien, C. E.
Knoder, G. L. Pearson, and C. Talbot for constructive suggestions and comments. While
many individuals discussed the topic with and in other ways assisted us, the report and
recommendations were prepared by the Conservation Committee and represent our
collective position.

LITERATURE CITED

Anderson, W. 1957. Waterfowl nesting study in Sacramento Valley, California, 1955.
Calif. Fish Game 43:71-90.

Burgess, H. H., H. H. Prince, and 1). L. Trauger. 1965. Blue-winged Teal nesting
success as related to land use. J. Wildl. Manage. 29:89-95.

Glover, F. A. 1956. Nesting and production of the Blue-winged Teal {Anas discors
Linnaeus) in northwest Iowa. J. Wildl. Manage. 20:28-46.

Leopold, A. S., C. Cottam, I. McT. Cowan, 1. N. Gabrielson, and T. L. Kimball.
1968. The National Wildlife Refuge System. Trans. North Am. Wildl. and Nat.
Resources Conf. 33:30-54.

Napier, L. D. 1974. Duck nesting success on the Stillwater Marsh. U.S. Fish and
Wildl. Serv. Unpubl. Rept.

SciiRANCK, B. W. 1972. Waterfowl nest cover and some predation relationships. J.
Wildl. Manage. 36:182-186.

U.S. Fish and Wildlife Service. 1976a. Environmental Impact Assessment. Effect of
boating on management of Ruby Lake National Wildlife Refuge. Dept. Interior.

. 1976b. Final Environmental Statement. Operation of the National Wildlife

Refuge System. Dept. Interior.

VoiGHT, W., Jr. 1976. Public grazing lands, use and misuse by industry and govern-
ment. Rutgers Univ. Press, New Brunswick, New Jersey.

Weller, M. W., B. H. Wingfield, and J. P. Low. 1958. Effects of habitat deterioration
on bird populations of a small Utah marsh. Condor 60:220-226.

Conservation Committee

Clait E. Braun, Chairman
Keith W. Harmon
Jerome A. Jackson
Carroll D. Littlefield

Wilson Bull., 90(2), 1978, pp. 322-325

ORNITHOLOGICAL LITERATURE

Rails of the World. By S. Dillon Ripley, illus. by J. Fenwick Lansdowne. Chapter
on fossil species by Storrs L. Olson. David R. Godine, Boston, 1977:xxd- 406 pp., 41
color plates, 35 black-and-white plates, 17 maps; size 10 X 14 in. $75.00. — Dillon Ripley
has been interested in rails since his earliest expeditions abroad many years ago, and
bas long wanted to write a monograph on the family. He was both inspired and chal-
lenged by Alfred Newton’s remarks in tbe Dictionary of Birds <1893) that the rails are
an enigmatic and little known family for which a monograph would prove rewarding
and valuable. Ripley further explained his purpose in a meeting at the Peabody Museum
in 1977, as quoted in Discovery 12(3) :44 (1977), that he had become certain over the
years “that if I ever could write a monograph on the rails, not only had one never
been written before, but the likelihood was that one would never be written again, and
that almost anything I managed to say about the rails was likely not to be very strongly
corrected in the future, so that I had clear sailing by taking on a family that is as
obscure, and little known, and enigmatic as the rails.”

The result of Ripley’s endeavors is a handsome monograph that certainly can take
its place in the parade of fine bird books of recent years. This is a book in the grand
manner; large, lavishly illustrated, and printed on fine paper with wide margins. It
combines the attributes of a “coffee talile” book that is a pleasure to browse through,
and a utilitarian work of inestimable value to all who are interested in rails.

Fenwick Lansdowne is an accomplished artist whose rails here are generally success-
ful and attractive, though a few have the wrong shape or pose. The Spot-flanked
Gallinule is drawn to look like a gallinule, which it is, but in the field it looks like an
oversized Porzana, and its bill is much brighter than pictured here. One irritating thing
about the paintings is that the birds are not drawn to scale. On the first plate, for in-
stance, the giant bird in the middle of the picture is Eulabaeornis castaneoventris, length i
11 in., while the small one crouching at the top left corner is Himantornis, a hefty
17-inch bird. There are many other examples of this lack of scaling. The grouping is ■
sometimes curious. Two Canirallus are on Plate 2 with Wallace’s Rail, while the third 1
is on Plate 3; why not put all 3 together? Perhaps the artist did the plates before the
taxonomist started playing around with the genera. Another irritation is that while the ;
text is keyed to the plates, the reverse is not true. The plate captions should have con-
tained page references to the species accounts. Despite these annoyances the plates
are a fine addition to the book and make a large contribution to its success.

The book is divided into an introduction and 5 parts. In the introduction Ripley gives
a history of the discovery of rails, written with a charming antiquarian flavor, and?
describes his own early involvement with the family. His pleasant writing style makes i
for easy reading.

The acknowledgments come at the end of the introduction. From the vague generaliza-
tion that “in recent years Gorman Bond and Michel Desfayes have undertaken further to I
make this whole work possible for me,” one would never guess that what these 2 really !
did was to research and write up much of the data for the present work. Ripley should ,
give pro])er, specific credit where due.

Part 1, The Characteristics of Rails, has sections on anatomy, classification, locomotion, ■
flight, nesting, voice, duetting, display, fighting, and various aspects of the relationships j
between rails and man, with a long section on hunting. In the classification on p. 5, only j
3 of the 8 promised gruiform suborders are listed; the other 5 were somehow left out. j

322

ORNITHOLO(;iCAL LITERATURE

323

The section on classification deserves some comment, mainly because of Ripley’s
radical revision of rail genera. His classification can be compared to those of previous
authors as follows:

Genera

Species

Sharpe 0894), Catalogue of the Birds in the British Museum, vol. 23.

165

Peters <1934), Check-List of Birds of the World, vol. 2.

138

Berndt & Meise 0960), Naturgeschichte der Vogel, vol. 2.

138

Fisher & Peterson 0964), The World of Birds.

130

Olson 0973), A Classification of the Rallidae, Wilson Bull. 85:381-416

Ripley (1977), Rails of the World.

129

It can be seen that while Ripley has made few changes at the species level, the
carnage among the genera has been great. While few would deny that the time had
come to do away with many of the old monotypic genera, one wonders whether such
wholesale slaughter is really justified. There is no doubt in Ripley’s mind that it is,
for he says ( p. 6) ‘T find few of the reasons advanced for maintaining additional genera
compelling or cogent. A single question . . . seems to me open to arguments of taxonomic
taste or discretion.” In other words there is only one correct treatment — his own. At-
tention, rail taxonomists! The final word has now^ been spoken. You should be looking
for work in other fields. Ripley continues: “For the rest, the preservation of so many
genera of rails up until now has been a question of waiting for monographers to address
themselves to this neglected family.” This is true, but the reader should not get the
impression that Ripley w^as the pioneer monographer who made the break with tradition.
That pioneer was Storrs Olson, who reduced the number of genera to 35. Incredibly,
Olson’s classification is not even mentioned in the list of previous classifications on p. 6.
Here is another instance of proper credit not being given.

In part 3, Evolution and Speciation, Ripley gives his reasons for merging various
genera. Arguments will probably continue forever as to what constitutes a valid generic
character, and every taxonomist is entitled to his own opinion, the only requirement
being that he be consistent in his use of characters. The same character must not be
used for 2 opposing functions. While many of Ripley’s generic limits seem reasonable,
he is sometimes guilty of this no-no. Thus, plumage similarities are the basis for merging
^eocrex in Porzana and Aramides in Eulahaeornis, while plumage dissimilarities are
ignored in uniting the African genus Sarothrura with the totally dissimilar New World
genera Coturnicops and Micropygia. This arrangement is in any case geographic non-
sense; and no-one who has ever heard the voices of both the Yellow Rail and Sarothrura
spp. would ever dream of putting them in the same genus.

Part 2, The Distribution of Rails, is the most successful of the introductory sections,
concentrating on the causes of distriI)Ution and covering subjects like adaptation and
dispersal. Among other topics Ripley discusses the reasons for the success of rails
as colonizers, and shows how their reluctance to fly and their rapid evolution of flight-
lessness might paradoxically help rather than hinder colonization. This is a thought-
provoking and well-written section.

Part 4, The Species of Rails, contains the species accounts. This is the real meat of
the book and its most valuable contribution. It is prefaced by a key to the genera, while
in the text the genera have species keys and the species have subspecies keys. Each
species account contains a brief overall description of the bird, taxonomic comments, and

324

THE WILSON BULLETIN • Vol. 90, No. 2, June 1978

a discussion of habitat, voice, food, and breeding. Each subspecies account contains a
detailed description, measurements, distribution, and status. There are only 17 maps,
and these are reserved for polytypic species with the most complex distributions. The
maps are well-drawn, clear, and easy to read. There is a tremendous amount of in-
formation here, and the researchers evidently did a good job with the literature since
the text is freely littered with citations. While excellent as regards “museum” type
information (description, measurements, subspecific characters, distribution, etc.), the
text would have benefitted from more first-hand field experience of the writers. Lacking
this an attempt should have been made to contact other workers in the field and obtain
unpublished information from them. In the account of Rallina tricolor, for example, out
of nearly 2 full pages of text there are just 3 lines on life history as follows: “From
the little that is knowm of its life history it appears to share the ralline habits of secrecy
and stealth, damp habitat, and crepuscular calling and feeding preferences.” You could
make this same statement about most of the rails of the wmrld and get away with it.

Yet there is specific information about the bird. In 1974 Mrs. H. B. Gill, a well-known

Australian amateur ornithologist, took me to see a pair of Rallina tricolor inhabiting a
stream near her house. I taped the calls and played them back, w^hich brought the
birds into view. But I was not witnessing something new to science. Mrs. Gill had kept

the birds under observation for some time and had already shown them to a number

of visitors. Here was a paragraph of information on the species just w^aiting to be tapped.
Also, there are only 2 lines on status in the entire account: “By no means rare in

Queensland (E. P. Ramsay in Mathew^s, 1911).” Surely more is known about its
status than a single remark in a book written 66 years earlier. Again, the status of
Canirallus k. kioloides is given as “Especially common on the narrow coastal plain at
Maroantsetra IRand, 1936).” That may have been true in Rand’s day, hut today most
of the vegetation there has been cut dowm and does not harbor Canirallus, as Ripley
could have discovered by checking with recent workers in ^Madagascar. Rouget’s Rail,
Rougetius rougetii is said to “keep hidden in thick vegetation during the day,” and
appear at dusk and dawn, another safe statement about almost any rail, but this one is a
maverick and comes boldly into the open far from cover during daylight hours. I have
sat in my car on the main highway north of Addis Ababa and w'atched Rougetius feeding
on bare mud by a small pond, scarcely blinking as trucks and buses hurried by. The
above instances only indicate a certain lack of field experience wdth the birds concerned
and a lack of depth in research, and should not obscure the overall value of the in-
formation presented.

While the species accounts fulfill their function of drawing together information from |
every source, there is also new information of interest. Ripley has stuck his neck out
and lumped King and Clapper rails, a treatment that will annoy many while pleasing

others, hut Ripley has bolstered his case with a lot of data, including personal com-

munications, and has produced a reasonable argument. Again, there is a lot of informa- j
tion on the Galapagos Rail, Laterallus spilonotus, derived to a great extent, it would |
seem, from an unpublished MS by Alan Franklin and Deborah and David Clark, who j
studied the bird in the field. Strangely, there are no taxonomic comments on this bird, |

which has hitherto been considered a subspecies of the Black Rail, Laterallus jarnaicensis. j

While this separation may well he justified. I think we are owed an explanation.

Part 5, A Synopsis of the Fossil Rallidae by Storrs L. Olson, is worthy of high praise.

It is much more than a simple catalog. Each species is written up under the following
headings: Holotype, Horizon, Locality, Material, Illustrations, and Remarks, the latter
often containing considerable discussion. The chapter is copiously illustrated wdth

ORNITHOLOGICAL LITERATURE

325

photographs and drawings of bones. This is a most valuable contribution by one of the
world’s leading authorities on the subject.

In conclusion, despite many faults this is a worthwhile and valuable publication. To
have put all the rails of the world between two covers, complete with illustrations, is a
praiseworthy achievement in itself. The price is steep but not out of line with today’s
incredible book prices, nor unreasonable considering the quality of production and
number of illustrations. For all who can afford it, I can certainly recommend the
book. — Stuart Keith.

Wilson Bull., 90(2), 1978, pp. 325-327

The Audubon Society Field Guide to North American Birds. Eastern Region.
By John Bull and John Farrand, Jr., Alfred A. Knopf, produced by Chanticleer Press,
New York, 1977:775 pp., 584 color photos. $7.95.^ — ^A bird guide that tops the New
York Times Book Review best seller list for paperbacks (which it is not) for several
weeks, and which remains on that list for several months deserves more than passing
mention. The dust jacket claims the book to be “a revolutionary field guide. Unique on
four counts.” The supposed novel ideas are to illustrate the various species with
color photographs rather than paintings, a “visual organization” of these photographs
by color and shape, and a text arranged by habitat. None of these ideas is really new.
The argument as to whether a photograph or a painting is the best means to present
a typical representation of a bird is an old one (see a review by R. Mengel, Auk 72:
308-310, 1955), and in this case our authors, both connected with the American Museum
of Natural History, come out wholeheartedly for photographs, “every artist rendering
of a bird is his interpretation whereas a good photograph captures the natural color
and stance of birds as you usually see them” (p. 12). The operative word is, of course,
“good.”

The extent of the participation of the Audubon Society (National) in the preparation
of the book is nowhere made clear, although Susan Rayfield of the staff of Audubon
is listed as “Project Editor” and is given credit for the development of the “Visual Key.”
I understand that it was she who performed the herculean task of rounding up the
photographs used. In some ways this guide shows signs of lineal descent from the
“Audubon Guides” written by Dick Pough a generation ago. Much the same geographical
range is covered (everything east of the Rocky Mountains), and the present work
attempts as did Pough to provide information beyond simple identification matters.

There are 584 color photographs covering 456 species. Most species are represented
by spring males only, but 122 species have a second illustration, usually of a female or
of another plumage. Twenty-six species are described but not included in the photo-
graphs, although a very small drawing of these species accompanies the text.

The photographs are clumped in the first half of the liook, arranged in groups of
similar shape, with small silhouettes serving as location guides. The passerines are ar-
ranged hy color. The 2 photographs of dimorphic species are thus often widely separated.
Arrangement by color also meant that some arbitrary decisions had to be made as to
under what color to include a given species. These decisions were sometimes rather
unwise, as for example the predominately gray and yellow Western Kingbird is listed
under green birds. The text accounts of the species are arranged under 12 categories of
habitat. Here again arbitrary and not always fortunate decisions had to be made. The
Great Horned Owl, which must nest in almost any haliitat, is listed under “Coniferous
Forest,” the Red-headed Woodpecker is listed under “Grasslands” and inland birders

326

THE WILSON BLLLETIN • Vol. 90, No. 2, June 1978

will be somewhat at a loss to find many familar ducks listed under “Seashores.” Usually
these inconsistencies are set to rights in the text description of habitat, however.

Several appendices are included: thumbnail sketches of the avian families occurring
in North America, a brief essay on bird watching, a glossary, a list of endangered
species, and a list of photographic credits. The arrangement of this latter list makes
it very difficult to find out which photographer gets credit for a particular picture.

The color pictures deserve critical comment. Color reproduction on a mass basis
has come a long way in recent years and the job done here is an outstanding one. At
the moment I know of no other collection of beautiful color photographs of birds
that can be obtained so cheaply. Most of the pictures are excellent, although a few
show unusual or even distorted poses. I could detect very few color distortions in my
copy. The delicate sky-blue of the Mountain Bluebird and the Lazuli Bunting have
reproduced to look more like the harsher blue of the Indigo Bunting and some of the
other blues are slightly off. None of the plates in my copy is out of register.

A few of the pictures are bad, however, with the nadir being possibly the Philadelphia
Vireo (451), and a number of others, while not bad as photographs, fail to show field
marks of aid in identification. Examination of some of the pictures used, as well as the
list of species not figured would indicate that even today some of our North American
birds have not been adequately photographed. The Philadelphia Vireo may fall in this
category. Unfortunately, there are some errors in identification. Plate 387 which is
supposed to represent the presumably rarely photographed Black-headed Oriole (Icterus
graduacauda) is actually one of the Old World orioles, possibly Oriolus larvatus which
is also known as the Black-headed Oriole. Plate 268 is a female Spruce Grouse rather
than the labeled Ruffed Grouse and Plate 37 appears to be a Western Gull rather than
the indicated Herring Gull.

A definite hazard of using photographs for a field guide comes from the fact that many
species are most commonly photographed on the nest or at least on the breeding grounds,
and for some species the full breeding plumage is rarely seen by most bird watchers.
Thus, the very fine picture of Sabine’s Gull (44) at the nest would be of little help
in identifying a Avinter-plumaged bird off the Maryland coast. Similarly, phalaropes in
breeding plumage and alcids in close view are seldom seen by birders. Perhaps the most
flagrant example of this is the Ruff (Plate 214) in its elaborate breeding plumage,
photographed in full display, a charming and interesting picture but one bearing
practically no resemblance to any Ruff ever seen on this side of the Atlantic.

From many years of experience in showing beginning bird watchers live birds in the
hand at banding stations, I have become aware that often the tyro is overwhelmed by the
detail in the plumage patterns. This hazard is also prevalent in the current guide. The
duck pictures are by and large the most beautiful I have ever seen, but the detail of the
feather patterns, the fine vermiculations, and the play of colors evident in these pictures
are certainly not apparent when viewing ducks across an expanse of choppy water on a
rainy, windy day.

All of the birds in the photos are reproduced to the same size, and even though a
measurement of length is given for each species I fear that incongruous size relations
will hinder rather than help the tyro. Some of the backgrounds in the pictures are un-
fortunate, the most ludicrous being a Chimney Swift posed awkwardly clinging to the
side of a smooth-barked tree.

We are on surer ground when we consider the text. Both of the authors know their
bird identification and they do a good job of describing the salient features for identifica-
tion in the brief paragraphs allotted to this point. A few inconsistencies do occur. For
example, on page 651 we are told that the Mourning Warbler lacks an eyering, but on

ORNITHOLOGICAL LITERATURE

327

page 706 we learn that it does have one. To add to the confusion the male in the photo-
graph (372) does not have one hut the female in the same photo does. In fact, many
Mourning Warhlers, particularly fall females do have at least a partial eyering, a char-
acter seldom mentioned in any field guide. An inconsistency of another sort comes when
we are informed on page 658 that the Black-crested Titmouse is included under the
Tufted Titmouse, hut on page 613 it is considered a full species.

Besides the descriptions, the text entries for each species include a very sketchy
description of the voice, and range descriptions that are so abbreviated as to be misleading
in some cases, particularly those northern species that nest south along the Ap-
palachians. There are also habitat descriptions that are generally adequate and as
mentioned above go a long way towards correcting the misleading categorization by
habitat, and a brief description of the nesting habits. At the end of each species account
there is a short paragraph presenting some additional information about the species,
ranging from the trivial, through the self-evident, to some worthwhile and interesting in-
formation. As with the habitat categories there appears to he some slight bias towards
things of interest to coastal bird watching and some of the remarks fail to apply to the
given species when found inland.

How' then does this hook in the final analysis stack up as a field guide? Regretfully
I must decide, “Not well, at all.” It will he a rare novice bird watcher who can identify
any hut the most obvious species (usually only males) with this guide. Such things as
the grassland sparrows, the shorehirds, and particularly the raptors will he very difficult.
The fall warblers would he impossible.

The classic advice to beginning bird watchers is to equip themselves with a good field
guide and also a set of good colored pictures of the birds. I suggest that for the beginner
this hook is a very good one for the second purpose to go along with one of the better
guides. It is true that he will have a hard time finding any particular species except
by way of the index, hut he can enjoy the pictures. For the experienced birder I would
suggest that the price is right for a good set of color bird pictures. Indeed the ex-
perienced field birder might well profit by careful study of the detailed plumage char-
acters shown in some species. On page 12 the authors say that the hook is “ — meant to
be a delight to look at. . . .” By and large this objective has been attained even if
the book falls short in other respects. — George A. Hall.

Wilson Bull., 90(2), 1978, p. 327

Statistical Inference from Band Recovery Data: A Hand Book. By Cavell

Brownie, David R. Anderson, Kenneth P. Burnham. & Douglas S. Robson. U.S. Dept, of
the Interior. Fish & Wildlife Service, Resource Publication No. 131, Washington, DC, 1978:
212 pp. No price given.

Wilson Bull., 90(2), 1978, pp. 327-328

The Pheasants of the World, 2nd Edition. By Jean Delacour. Spur Publications,
Saiga Publishing Co., Ltd., Hindhead, Surrey, England. 1977: 395 pp., 33 plates, 21
text figures. £18. — This is an updated reissue of the standard work on pheasants, first
published in 1951. The original text is reprinted without revision, hut more recent in-
formation is included in addenda to the various sections. After a general introductory

328

THE WILSON BULLETIN • Vol. 90, No. 2, June 1978

account, each species is considered individually in terms of appearance, habits, dis-
tribution, and especially in relation to care and breeding in captivity.

In addition to numerous range maps, the book is handsomely illustrated with 32
plates painted by J. C. Harrison, half of them in color. Some of the quality of the
original color plates has been lost in reprinting. There is a reduction in sharpness from
the first edition, and some differences in color tones. Many plates have acquired a
distinct greenish tinge. New to this addition is a frontispiece, an attractive painting of a
male Himalayan Blood Pheasant by R. David Dighy.

Anyone interested in pheasants, both in nature and in captivity, will welcome the re-
appearance and revision of this authoritative and visually pleasing work. — Robert J.
Raikow.

Wihon Bull, 90(2), 1978, pp. 328-329

Species Relationships in the Avian Genus Aimophila. By Larry L. Wolf. Ornitho-
logical Monographs No. 23, American Ornithologists’ Union, 1977: viii + 220 pp., 10 pL,
17 figs., 36 tables, 1 long play record. $12.00 ($10.50 to A.O.U. members). — Even Robert
Ridgway, who in 1901 established the genus Aimophila in its present form, felt that it
was “a very heterogeneous and probably unnatural genus.” Lack of information about
the species in this genus has hampered attempts at arranging them into related groups,
and a similar lack of information about many related emberizine finches has made
intergeneric studies all but impossible. Dr. Wolf has added enormously to our under-
standing of relationships within Aimophila, but comparable studies of several possibly
related genera are still needed, as are similar studies of the two South American species,
Aimophila strigiceps and A. (Rhynchospiza) stolzmanni, which Wolf was unable to
attempt.

This monograph contains a wide range of information on the 12 North and Middle
American aimophilas, including geographic and ecological distribution, molts and
plumages, territoriality and pair bond, foraging and food, vocalizations, breeding seasons,
nest structure, egg color, external morphology, and skeletal characters. There is no list
of specimens examined, hut sample sizes on the tables indicate that more than 1100 study
skins and 340 skeletons were examined, and weight data on 667 specimens were included.

Based on his analysis Wolf divides the genus into 4 complexes: the Haemophila

complex ( ruficauda, sumichrasti, humeralis, mysticalis, and carpalis) having “radiated
in the lowland scrub forests of western Mexico and the Pacific lowlands of Central
America”; a ruficeps complex (ruficeps, riifescens. and notosticta) having radiated in
“pine-oak woodland of Mexico and Central America,” a botterii complex (aestivalis,
botterii, and cassini) in “weedy, open country of Middle America and United States”;
and quinquestriata of “dense deciduous woodland” of northwestern Mexico.

Not considering the South American species detracts somewhat from the zoogeographic
analysis. Aimophila strigiceps of northern Argentina and Paraguay resembles A.
sumichrasti in plumage and inhabits brushy fields. A. stolzmanni of southwestern
Ecuador and northwestern Peru differs from other aimophilas in its large bill, but
appears closest to the Haemophila complex in plumage, and like strigiceps, inhabits brushy
areas. Eliminating these species from possible membership in the other 3 complexes
strengthens the zoogeographic conclusions regarding these complexes, while their possible
inclusion in the Haemophila complex suggests that the conclusions regarding the evolu-
tion and distribution of that group will need expansion and modification.

ORNITHOLOGICAL LITERATURE

329

The 12-inch record, on which songs of all 12 species and “chatters” of 5 are recorded
will prove interesting and useful to ornithologists but a headache to librarians. On the
whole, the vocalizations on the record well complement those shown on the sonagrams,
but cross referencing and editing of the data could he greatly improved. There are no
references in the text to the vocalizations on the record, and not all song types on the
record are shown in the sonagrams. Nor is there any indication that any sonagram was
taken from any song on the disc, although some may have been. Finally, while there
are lists of figures and tables in the introductory material, there is no comparable list
of the plates showing which vocalizations of which species are represented on each.
In the absence of an index, such a list would have been very useful.

This monograph includes a wealth of information on a complex group of finches and
is particularly valuable in showing how different kinds of information may be pooled to
produce a good understanding of interrelationships wdthin a genus of birds. If com-
parable studies of such genera as Melozone, Oriturus, Pipilo, and Chondestes are pursued,
a better understanding of the relationships among these genera and the complexes within
Aimophila will follow. — Robert W. Storer.

Wilson Bull, 90(2), 1978, pp. 329-330

Guide to The Young of European Precocial Birds. By Jon Fjeldsa, illus. by the
author. Skarv Nature Publications, Strandgarden, DK-3220 Tisvildeleje, Denmark, 1977:
285 pp., 39 color plates, 70 text figs., 1 photo. Danish Kroner 200 lapprox. $33). — This
is the first guide devoted exclusively to downy young for any part of the globe, and
emphasizes 180 species from Europe and Greenland. More than 70 of these species also
breed in North America excluding Greenland. In addition to explaining and illustrating
characters for identifying downies, the volume includes comments on their ecology and
systematics.

Ornithologists have often neglected downies. Collectors often fail to preserve them,
and perhaps for this reason, taxonomists have frequently slighted them. As Fjeldsa
points out, downies in the field are preferably identified by their own features rather
than by the adults present as is often done. Where similar species breed in the same
locality, identification is particularly challenging. For Europe and Greenland Fjeldsa
gives characters to identify all precocial downies to species, even for such difficult
groups as the gulls, whose downies have often been thought to be indistinguishable.
Fjeldsa has handled more than 3700 live and preserved downies, an impressive total in
view of their scarcity in many collections.

A general account of the biology of precocial birds (pp. 19-23) surveys relationships
between nidifugous habits and embryonic maturation of young, nest sites, clutch sizes,
adult foraging, and other ecological features. A section on morphology and changes in
proportions during growth (pp. 24—30) includes a brief review' of taxonomic variations
in the structure and appearance of natal plumages. Natal downs exhibit a rachis in only
a few avian families including certain ratites, tinamous, some Galliformes, and the
Anatidae. As the natal downs of flamingos lack a rachis, the often noted similar appear-
ance of natal plumage in flamingos and swans is superficial. Fjeldsa also examines the
relationship between habitat and the amount of pattern in downy plumages.

A section on banding downies (pp. 31-34) describes the reshaping of bands to match
the cross sectional shape of tarsi. Many chicks too small for banding with conventional

330

THE WILSON BULLETIN • Vol. 90, No. 2, June 1978

bands that would slip off the leg can he marked l)y using such hands with a plasticine
inner lining that gradually wears away as the leg grows.

A major part of the hook contains family and species accounts. English common
names follow Euroj)ean rather than American usage with synonyms given for each spe-
cies in Danish, German, Dutch, and French. In a typical species account Fjeldsa de-
scribes the downy young in detail, referring to one or more illustrations. Characters for
distinguishing similar species are emphasized, and the author notes and illustrates indi-
vidual variations in downy plumage for especially variable species. Also listed are body-
length and weight at hatching, time from hatching to flight of young, maximal brood
sizes, hand size, breeding range in Europe and Greenland, habitat, and time of year at
which downies occur. Included are loons, grebes, a flamingo, swans, geese, ducks,
grouse, phasianids, a button quail, bustards, rails, a variety of Charadriiformes, and sand-
grouse. In accord with Fjeldsa’s taxonomic views loons and sandgrouse are placed in
the Charadriiformes.

Brief comments on evolutionary relationships are included, and downies of a number
of ncnEuropean species are illustrated and or discussed briefly. Certain taxonomic dif-
ferences in downy patterns are termed “nonadaptive” fp. 12), hut would perhaps be
better characterized as alternative kinds of adaptation. “Morphocline” phyletic dia-
grams indicate possible evolutionary affinities within the grebes, Anatidae, grouse, and
Charadriiformes, hut the evidence supporting these diagrams is not fully presented. In
one such diagram eiders are placed in an unconventional arrangement between golden-
eyes and mergansers. In the same figure, Fjeldsa’s sketch of a downy Anseranas does
not agree with the correct description and illustration provided by Delacour (Waterfowl
of the World, 4:327 and Plate V, Country Life, London, 1964).

Imperfections of this volume appear minor relative to its considerable merits. Neither
a number of spelling errors nor an occasional sentence with peculiar wording detract
seriously from the overall presentation. A statement ( p. 27) that in certain taxa a plum-
age of preplumulae is pushed out by a plumage of prepennae needs clarification; it is
difficult to visualize how this could occur. Fjeldsa uses the term mesoptile but unfor-
tunately does not explain how mesoptile plumages as a category differ from other kinds.
Regrettably the text lacks literature citations, though three pages list the consulted pub-
lications; a reader wishing to check particular statements could have difficulty finding
the sources.

For identifying downy chicks of European precocial species I know of no other book
equaling this volume. Many of the author’s illustrations are outstanding, and some may-
purchase the hook for no other reason. The volume is attractively produced, sturdily-
bound. and fits readily in a field knapsack. It is an appropriate addition to any library
with major ornithological holdings, and luologists with special interests in the downy-
young of any of the families or species considered will want access to it. — George A.
Clark, Jr.

Wilson Bull, 90(2), 1978, pp. 330-332

The Courtship of Birds. By Hilda Simon. Dodd, Mead & Co.. New York, 1977;
190 pp., 54 color illustrations by the author. S12.95 — This hook describes the diverse
ways that '“birds woo and win their mate.” Apparently- intended for amateur ornithol-
ogists, lieginning students of birds and behavior, and nature lovers, it could serve as a
good introduction to courtship behavior for any beginning student from junior high school

ORNITHOLOGICAL LITERATURE

331

on. The text is divided into 5 chapters: Patterns of Courtship Behavior, Plumage

Pageantry, Display Acrobatics, Wooing by Work, and Mutual Courtship,

Each chapter begins with a brief poetic description of the courtship of a species
typical of the category being discussed. This serves to interest the reader and to introduce
the topic. The first chapter describes the general courtship patterns of birds, and provides
an introduction for the rest of the text. Simon discusses species recognition, imprinting,
social facilitation (not so named), the effect of light, and aerial courtship. The chapter
on Plumage Pageantry discusses how birds use feathers in courtship, and includes de-
scriptions of birds of paradise, lyrebirds, and the Great Argus Pheasant. In Display
Acrobatics the author describes courtship in grouse, prairie chickens, Ruff, bustard.
Ostrich, and manakins. Wooing by Work understandably includes descriptions of bower-
birds, frigate birds, penguins, tits, wrens, and weavers. In the chapter on Mutual Court-
ship Simon describes Whooping Cranes, swans, geese, ducks, grebes. Jackdaws, and
Laughing Gulls. As is obvious from the above list of species, she covers a wide range
of species representative of birds in general. As when many non-ornithologists describe
avian courtship, she concentrates on the spectacular, showy species whose antics and be-
havior fascinate the naturalist in all of us. However, sufficient space is not devoted
to the majority of species whose monogamous courtship patterns are less spectacular.
This treatment will surely give the naive reader an inaccurate picture of the number
of species with spectacular courtship patterns. This solitary failing in the book can
be forgiven since its intent is to excite interest, and this it does through a lively writing
style and delightful illustrations.

Simon’s writing is clear and pleasant. She treats such complex behavioral concepts as
sexual dimorphism, imprinting, leks, and anthropomorphism well. Her discussion of the
role of learning in imprinting, although brief, avoids the usual pitfalls, and she manages
to avoid condescension while carefully defining all technical terms. Her writing style
is expository, yet poetic. Occasionally her sentences become overly long, sometimes taking
an entire paragraph, which may obscure tbe meaning in some cases.

I found the book to be generally accurate, l)earing in mind that the descriptions are
brief. However a few points need correction in a future edition. Although a bird may well
sing to “encourage his mate” ( p. 15), we have no way of knowing this. For a further
discussion of such topics readers should refer to D. Griffin’s recent book The Question
of Animal Awareness. Altbough ethologists argue about the role of males and females,
we still do not refer to males as asserting their dominance, to females as childlike
( p. 66), or to a female as seeing the “error of her ways” ( p. 128). I know of no
ornithologists who “expressed a vague irritation” as they described “the seeming in-
difference of the peahen when faced” with the splendor of the male’s display fp. 70).
The role of males displaying on a lek may be socially facilitating as well as for establish-
ing a hierarchy. From my own perspective, the description of Laughing Gull courtship
needed editing. I know of no research that indicates that males and females cannot
recognize the sex of another Laughing Gull. The charge of a male is to repel intruders,
regardless of their sex! The female’s appeasement behavior may initiate courtship,
but it is not an indication that the male did not perceive her as a female. Additionally,
food-begging is not only symbolic in this species, as males do indeed regurgitate food
to their mates. Generally, however, the descriptions are accurate, sometimes misleading
only because of their brevity.

The color illustrations by tbe author sui)plement the descriptions, giving the reader a
picture both of the bird and the display or behavior being described. Most illustrations
are accurate, although the color on some could be slightly imi)roved. For the non-

332

THE WILSON BULLETIN • VoL 90, No. 2, June 1978

taxonomist it would help to identify the species and behavior in the illustrations, al-
though a legend for each illustration does appear at the beginning of the book.

In general, this book is well-written, interesting, fairly accurate, and well-illustrated.
I recommend it highly for amateur birders, high-school students, and anyone who enjoys
nature. Additionally, it might be very instructive reading some evening for beginning
undergraduate ornithology and behavior students.— Joanna Burger.

Wilson Bull, 90(2), 1978, pp. 332-334

Manual of Neotropical Birds, Vol. 1. By Emmet R. Blake. The University of
Chicago Press, Chicago and London, 1977: 674 pp., 12 plates (4 in color), 67 wash
drawings, numerous range maps. $50.00. — South America has been called the “bird
continent,” as the variety of its feathered inhabitants surpasses that of all other tropical
land areas of the world. However, a detailed descriptive account of the avifauna in its
entirety has never been published. The “Manual” is designed to fill this gap and is the
first of a projected series of 4 volumes. The work, once completed, wdll provide a
synthesis of basic data of “all species and subspecies of birds recorded from the main-
land of both Central and South America, the continental islands, and adjacent waters”
(excluding Mexico, the West Indies, the Galapagos and Falkland Islands). Following
the Wetmore sequence, the first volume includes mostly “coastal” and “water birds.”
Among the more typically South American families treated are the rheas, tinamous, and
cracids (curassows, guans, chachalacas) . A brief introductory and general section in-
cludes a useful synopsis of families (by K. C. Parkes). The main text provides keys by
families for all species to determine the bird “in the hand” and detailed descriptions of
species and of all subspecies as recognized by the author. Measurements are given (sample
size, range, mean) and the distribution is described and in most cases illustrated by
fairly small, yet highly instructive maps. No life history data are summarized, pre-
sumably for reasons of space, but references for each species serve as a guide to ecological
or more specific “biological” publications. Recent taxonomic sources followed by the
author are quoted under the family and under many genus headings. The sequence ]

of species in some families or genera and the generic allocation of certain species differ ^

from those used by de Schauensee in his reference lists of the South American avifauna I
( The Species of Birds of South America and their Distribution, Livingston Publ. Co., !
Narberth, PA, 1966, and A Guide to the Birds of South America, Livingston Publ. Co., !
Wynnewood, PA, 1970). Of the 600 species treated in Blake’s first volume of the \
“Manual” just over one third (225 species) are illustrated in full or the head only on 12 |

plates (4 in color) and in 67 wash drawings scattered through the text. The superb ;

plates are by G. Tudor (except one) who also contributed several excellent text illustra- j

tions. The majority of the latter and one plate are by R. V. Keane. Unless the number !
of plates is increased in future volumes, an inadequate illustrative coverage will be felt '
especially in the case of the very diverse passerine families. 1

Blake’s detailed systematic treatment of the Neotropical avifauna, summarizing and I
updating technical information scattered over a vast literature, will form a sound basis i
for future field investigators and will stimulate further ornithological studies in South j
and Central America. For the amateur it will be a dependable source of detailed in- I
formation on the feathered inhabitants of South America and Central America as far
as known today. Hopefully, the remaining volumes will be published without undue
delay.

ORNITHOLOGICAL LITERATURE

333

The influence of regional handbooks upon future ornithological research can hardly
be overestimated. Therefore, the author of such a work preferably presents the material
not only in the form of a summary and compilation of known data but, at the same time,
points out unknown aspects of the avifauna at various levels from individual species
and species groups to ecological communities. Generally speaking, the Neotropical
avifauna is poorly known. New species are still being described at a rate of several
per year. The life histories and the behavior of the majority of Neotropical birds have
not been studied and little is known on their seasonality and migratory behavior. Relevant
publications are listed in the “Manual” at the end of each species account. Nevertheless,
I hope brief statements like “Terrestrial forest bird . . .” or “. . . hunts for insect prey
in the canopy level. . . .” will characterize briefly the ecological stations of species in the
diverse Neotropical families to be treated in future volumes. As stated above, the
emphasis of the text is on plumage description, measurements and the distribution of all
species and their subspecies. The author does not discuss the theoretical basis for his
systematic treatment of South American birds and, in future volumes, hopefully will
point out more frequently open questions regarding the geographical variation and dis-
tribution of species and subspecies. “Subspecies” are often treated as if they were distinct
biological entities, their names and text printed in the same large letter type as that
used for the species (a smaller letter type for subspecies sections would be helpful in future
volumes). Many subspecies are clinally related, grading into each other over wide areas,
and their delimitation is highly subjective; or subspecies represent uniform populations in
geographic isolation from the main species range; in other instances subspecies or sub-
species groups (“megasubspecies”) meet along “hybrid zones.” These and other phenomena
of geographic variation and population structure and their relations to environmental
factors such as rainfall, seasonality of the climate, the existence of a network of broad
rivers, and isolation by mountain ranges need to be brought out for each of the South
American bird species (if the available data permit). In view of the differing nature
of geographic variation among Neotropical birds, a more flexible treatment of “subspecies”
might be considered for future volumes of the “Manual”: e.g., a brief introductory
section on general aspects of geographic variation would be useful for species with many
“subspecies”; several clinal forms might be discussed jointly as a subspecies group, if
necessary helped by tables for measurements, coloration or by maps to illustrate details
of distribution (with subspecies names mentioned in the text only). The rigid standard
treatment in this volume of species as if they were composed of ±;well “defined” sub-
species might he considered as typological. To be sure, it is of biological significance to
analyse the nature and cause of geographic variation in species populations, but to know
the names of subspecies is less important.

Species are also treated in the “Manual” as independent biological entities of equal
significance, although many different evolutionary levels are represented, from those
species which have barely reached reproductive isolation to distantly related sympatric
species. There are numerous South American species which exclude each other
geographically in fairly uniform ecological regions presumably as a result of competi-
tion. In many but not all of these cases geographical representatives may be combined
as allospecies of superspecies. Not a single pair of such parapatric species has been
studied along the contact zone to learn how parapatry due to competition ( ?) “functions”
in the field! For this reason it would be useful to map in detail the distribution of
species with “peripherally overlapping” or mutually exclusive ranges and to discuss
briefly species borders and relationships in an introductory section of each genus where
applicable. The useful if small scale maps included at the end of each family section

334

THE WILSON BULLETIN • Vol. 90, No. 2, June 1978

of the “Manual” depict the distribution of all species. Often several species ranges are
comliined on a map of all or part of South America as space permitted. It would be
useful to indicate in future volumes which of the allopatric or parapatric species could

be considered as superspecies or as species groups.

Summarizing, I feel that in the case of avian families with numerous related genera
and species tl.e reader of the “Manual” would welconte Itrief syntheses at various levels
which would help visualize the hiological significance of certain aspects of geograp iic
variation and distribution or which would point out certain problems of interspecific
relations. The linear treatment of species by necessity breaks the avifautia into seeming y
“independent” or “isolated” taxonomic units; the author could make an effort to
“rebuild” the fauna, at least partially, by including brief sections on systematic,
ecological or distributional aspects at the level of species, genera, and/or families
Hopefully, the author also plans for a future volume a section reviewing such general
topics as the regionally varying systematic composition of forest and nonforest avifaunas
resident versus migrant species, annual cycles among Neotropical birds, and historical
aspects of the differentiation of the South American avifauna.

In concluding I list a few comments concerning certain details of the text and the
maps. In coastal and montane species it might be advisable to indicate the range by a
heavy line or a series of dots following the coast line or a mountain slope, respectively,
to emphasize the linear extension of their ranges. A few minor corrections may e

listed; p. 29, key under C. strigulosus ( 9 ) south of the Amazon. The statement

of “many sightings of the Red-head tA. americana) in Costa Rica and central Panama
(p 2501 is based on a misunderstanding and does not refer to this species (E. Eisenmann,
pers. comm.). Daptrius americanus (p. 357, top) is distributed in Brazil south to Inot
oh Mato Grosso and S5o Paulo. Dr. J. O’Neill recently rediscovered Penelope alb.penms
(p 412) in northwestern Peru. Rhynchorlyx cinctus ( p. 453) ranges eastward m
northern Colombia to the Magdalena Valley iVolador; Westmore, Smiths Mtsc. Coll.
150 pt 1-332). Haffer (Puhl. Nuttall Ornithol. Club 14:106, 1974) considers the form
ochroptera as a subspecies of Psophia crepitans rather than of P. leucoptera Latemllus
exilis has also been collected in Amazonian Ecuador (Limoncocha; Pearson, Condor 77.
97 1975; this locality is not located in eastern Peru). Typographical errors are rare.

judging by this first volume, Blake’s “Manual” promises to become the basic de-
descriptive of Neotropical birds for many years to come, indispensible for professional
and amateur ornithologists alike. Jlrgen Haffer.

This issue of The Wilson Bulletin was published on 16 August 19^8.

The Wilson Bulletin

Editor* Jerome A. Jackson

Department of Biological Sciences
P.O. Drawer Z
Mississippi State University
Mississippi State, MS 39762

Editorial Assistants Bette J. Schardien Patricia Ramey

C. Dwight Cooley Martha Hays

Renne R. Lohoefener

Review Editor Robert Raikow Color Plate Editor William A. Lunk

Department of Life Sciences 865 North Wagner Road

University of Pittsburgh Ann Arbor, MI 48103

Pittsburgh, PA 15213

Suggestions to Authors

See Wilson Bulletin, 87:144, 1975 for more detailed “Suggestions to Authors.”
Manuscripts intended for publication in The Wilson Bulletin should be submitted in dupli-
cate, neatly typewritten, double-spaced, with at least 3 cm margins, and on one side only
of good quality white paper. Do not submit xerographic copies that are made on slick,
heavy paper. Tables should be typed on separate sheets, and should be narrow and deep
rather than wide and shallow. Follow the AOU Check-list (Fifth Edition, 1957) and
the 32nd Supplement (Auk, 90:411-419, 1973), insofar as scientific names of U.S.
and Canadian birds are concerned. Summaries of major papers should be brief but
quotable. Where fewer than 5 papers are cited, the citations may be included in the text.
All citations in “General Notes” should be included in the text. Follow carefully the style
used in this issue in listing the literature cited ; otherwise, follow the “CBE Style Manual”
(1972, AIBS). Photographs for illustrations should have good contrast and be on gloss
paper. Submit prints unmounted and attach to each a brief but adequate legend. Do not
write heavily on the backs of photographs. Diagrams and line drawings should be in black
ink and their lettering large enough to permit reduction. Original figures or photographs
submitted must be smaller than 22 X 28 cm. Alterations in copy after the type has been
set must be charged to the author.

Notice of Change of Address

The permanent mailing address of the Wilson Ornithological Society is: c/o The
Museum of Zoology, The University of Michigan, Ann Arbor, Michigan 48104. Persons
having business with any of the officers may address them at their various addresses
given on the back of the front cover, and all matters pertaining to the Bulletin should be
sent directly to the Editor.

* See Ornithological News, p. 308, for address for ms submission.

CONTENTS

A REVISION OF THE MEXICAN PICULUS (PICIDAe) COMPLEX LlUs F. Baptista 151

DISTRIBUTION, DENSITY, AND PRODUCTIVITY OF ACCIPITER HAWKS BREEDING IN OREGON

Richard T. Reynolds and Howard M. Wight 18!

SOCIAL AND FORAGING BEHAVIOR OF WARBLERS WINTERING IN PUERTO RICAN COASTAL SCRUB

William Post IS !

DDE RESIDUES AND EGGSHELL THINNING IN LOGGERHEAD SHRIKES

William L. Anderson and Ronald E. Duzan 21 ;[

TREE SPECIES USED BY BIRDS IN LOGGED AND UNLOGGED MIXED-CONIFEROUS FORESTS |

Kathleen E. Franzreb 25 1|

DOUBLE-BROODEDNESS IN PURPLE MARTINS IN TEXAS Charles R. Brown 221

FOOD OF NESTLING PURPLE MARTINS Helene Wolsh 24 j

REPRODUCTION AND NEST SITE SELECTION BY RED-WINGED BLACKBIRDS IN NORTH LOUISIANA

Bryan T. Brown and John W. Goertz 2C
THE RUFOUS-COLLARED SPARROW AS A HOST OF THE SHINY COWBIRD Rosendo M. Fraga 2" 1

GENERAL NOTES 1

SEXUAL SIMILARITY OF RED HEADED WOODPECKERS AND POSSIBLE EXPLANATIONS BASED j

ON FALL TERRITORIAL BEHAVIOR Lawrence Kilham 2(!

NOTES ON THE COURTSHIP BEHAVIOR OF BROWN-CAPPED ROSY FINCHES j

Paul Hendricks 2f |

EFFECTS OF NEST REMOVAL ON STARLING POPULATIONS j

H W Heusmann and Robert Bellville 2tl|

UNUSUAL INCUBATION BEHAVIOR IN BOBWHiTE George A. Hurst 2< !

A CATTLE EGRET-DEER MUTUALISM Marc R. Halley and Wayne D. Lord 2<:

A TEST OF SIGNIFICANCE FOR MAYFIELD’s METHOD OF CALCULATING NEST SUCCESS |

Douglas D. Dow

MIRROR IMAGE VERSUS CONSPECIFIC STIMULATION IN ADULT MALE ZEBRA FINCHES

Michael J. Ryan 2S

PROTOCALLIPHORA INFESTATION IN GREAT HORNED OWLS - Robert T. Bohm 2? ;

TERRITORIAL DEFENSE OF A NECTAR SOURCE BY A PALM WARBLER

Joseph M. Wunderle, Jr. 2<

RING-BILLED GULL PAIR WITH 2 NESTS - William E. Southern 2< ;

CLUTCH SIZE AND NEST PLACEMENT OF THE PIED-BILLED GREBE IN MANITOBA

Spencer G. Sealy 3(

NEST PLACEMENT IN SAGE THRASHERS ... Terrell D. G. Rich 3(

GREAT BLACK-BACKED GULLS BREEDING IN SALT MARSH IN NEW JERSEY

Joanna Burger 3< ;

^RESIDENT S PAGE

ORNITHOLOGICAL NEWS

CONSERVATION COMMITTEE REPORT

3<!

3<;

Bli

ORNITHOLOGICAL LITERATURE

Z.OOL..
’ . SF?R A RY

UEC 4 1978

VOL. 90, NO. 3

SEPTEMBER 1978 PAGES 335-478

The Wilson Ornithological Society
Founded December 3, 1888

Named after ALEXANDER WILSON, the first American Ornithologist. ®

President— Douglas A. James, Department of Zoology, University of Arkansas, Fayetteville,
Arkansas 72703.

First Vice-President— George A. Hall, Department of Chemistry, West Virginia Univer-
sity, Morgantown, W. Va, 26506.

Second Vice-President— Abbot S. Gaunt, Department of Zoology, Ohio State University,
Columbus, Ohio 43210.

Editor— Jerome A. Jackson, Department of Biological Sciences, P.O. Drawer Z Missis-
sippi State University, Mississippi State, Mississippi 39762. (See Ornithological
News, p. 308).

Secretary— Curtis S. Adkisson, Department of Biology, Virginia Polytechnic Institute
and State University, Blacksburg, Virginia 24061.

Treasurer— Ernest E. Hoover, 1044 Webster St., N.W., Grand Rapids, Michigan 49504.

Elected Council Members— James R. Karr (term expires 1979) ; Clait E. Braun (term
expires 1980); Sidney A. Gauthreaux, Jr. (term expires 1981).

Membership dues per calendar year are: Active, $10.00; Sustaining, $15.00;

Life memberships, $200 (payable in four installments) .

The Wilson Bulletin is sent to all members not in arrears for dues.

The Josselyn Van Tyne Memorial Library
The Josselyn Van Tyne Memorial Library of the Wilson Ornithological Society, housed
in the University of Michigan Museum of Zoology, was established in concurrence with
the University of Michigan in 1930. Until 1947 the Library was maintained entirely
by gifts and bequests of books, reprints, and ornithological magazines from mem ers
and friends of the Society. Now two members have generously established a fund fond
the purchase of new books; members and friends are invited to maintain the fund by
regular contribution, thus making available to all Society members the more important.!
new books on ornithology and related subjects. The fund will be administered by the.j
Library Committee, which will be happy to receive suggestions on the choice of books
to be added to the Library. William A. Lunk, University Museums, University of Michi-^«
gan, is Chairman of the Committee. The Library currently receives 104 periodicals as gifts 1
and in exchange for The Wilson Bulletin. With the usual exception of rare books, any.;
item in the Library may be borrowed by members of the Society and will be sent prepai
(by the University of Michigan) to any address in the United States, its possessions, or
Canada. Return postage is paid by the borrower. Inquiries and requests by borrowers,
as well as gifts of books, pamphlets, reprints, and magazines, should be addressed to
“The Josselyn Van Tyne Memorial Library, University of Michigan Museum of Zoology,^
Ann Arbor, Michigan.” Contributions to the New Book Fund should be sent to thelfi
Treasurer (small sums in stamps are acceptable). A complete index of the Library’s
holdings was printed in the September 1952 issue of The Wilson Bulletin and newly
acquired books are listed periodically.

The Wilson Bulletin

The official organ of the Wilson Ornithological Society published quarterly, in March September,

and December. The subscription price, both in the United States and elsewhere is "Sd be senJ to the
copies, S4.00. Subscriptions, changes of address and c kims for th^Treasurer Specia'

Treasurer. Most back issues of the Bulletin are available and may be ordered from the treasure P

SSr o' S' &ptsu.o up>ve„i„

Mississippi State, Mississippi 37962. Miner nffire

Second class postage paid at Mississippi State, Mississippi and at additional mailing office.

Allen Press, Inc., Lawrence, Kansas 66044 j

THE WILSON BULLETIN

A QUARTERLY MAGAZINE OF ORNITHOLOGY
Published by the Wilson Ornithological Society

VoL. 90, No. 3 September 1978 Pages 335-478

Wilson Bull, 90(3), 1978, pp. 335-345

AUTUMN BIRD CASUALTIES AT A NORTHWEST
FLORIDA TV TOWER: 1973-1975

Robert L. CpuAwford

Accounts of nocturnal accidents to migrating birds at tall, lighted struc-
tures are numerous ( Nisbet 1970) but most represent species listings cover-
ing 1 or 2 night kills. Tordoff and Mengel (1956), Goodpasture (1963a,b),
and Taylor ( 1972, 1973) showed that more detailed analyses of tower kills
contribute much to our understanding of many aspects of avian migration,
particularly to the differential migration of age and sex classes. Even so,
Johnson’s ( 1973) statement “. . . the temporal and geographic complexities
of migration . . . are only beginning to be understood” probably applies to
most migratory species. Many more data from different sites are needed.
This paper presents age and sex data for 3223 birds killed at the WGTV
tower in northwest Florida during the autumns ( August-November) of 1973-
1975. Comparisons with other studies, particularly to the one by Taylor and
Anderson (1973) in central peninsular Florida are made.

METHODS

Stoddard (1962), Stoddard and Noriis (1967), and Crawford (1974) gave the species,
numbers, and seasonal variation of birds killed at the WCTV tower on Tall Timbers
Research Station, Leon County, Florida, and complete desciiptions of the 308 m tower
and its 14 ha cleared site. In 1973. the grounds were checked for dead birds daily at
07:30 and the control of predators that eat the dead birds was essentially limited to
experiments (Crawford 1971, 1974). In 1974 and 197.5, the tower grounds were checked
at daybreak and a more rigorous program of predator control was instituted that in-
cluded both trapping and poisoning. In all years the birds were frozen soon after
collection and within 3 months of death were thawed and dissected to determine the
age and sex by examination of the skull and gonads (Miller 1946, Coodpasture 1963b).
I kept notes on plumage abnormalities, stomach contents, molt, and other physical as-
pects of most of the specimens.

RESULTS

Numbers and species recorded during the study. — In the autumn of 1973,
261 individuals of 57 species were collected. For the same period in 1974,

335

336

THE WILSON BULLETIN • VoL 90, No. 3, September 1978

1832 individuals of 87 species, and in 1975, 1771 individuals of 90 species
were found. The total for the 3 autumns was 3864 individuals of 109 species.
The disparity between the number of birds recorded in 1973 and subsequent
years is due primarily to the reintroduction of predator control at the tower.
Compared to 3 predators removed in 1973, 42 were removed from the tower
grounds in 1974-1975 and there was evidence that 28 others were eliminated.
Most of those handled were Virginia opossums (Didelphis virgmiarm), feral
domestie cats (Felis catus), and Great Horned Owls {Bubo virginianus) .

On 6 nights during the study period, more than 100 dead birds were re-
corded; 17 October 1973 (133), 5 September 1974 (134), 23 September
1974 (220), 17 October 1974 (971), 14 September 1975 (636), and 15
September 1975 (486). These are 66.7% of the grand total and all were
associated with the onset and passage of cold fronts.

Physical aspects of the tower casualties.— Plumage abnormalities were
rare. Partial albinism was observed in an adult 2 Yellow-rumped Warbler
(Dendroica coronata) and an unsexed adult Field Sparrow {Spizella pusilla) .
An adult 2 Bay-breasted Warbler [Dendroica caslanea) with an abnormally
pale-yellow head appeared to be partially xanthocroistic (Tall Timbers Res.
Sta. #3267). Molting individuals were also rare but nearly one-half of t e
Gray Catbirds (Dumetella carolinensis) examined, including all age and sex
classes, were in body molt. One immature <? Ovenbird (Seiurus aurocapdlus)

had sheathed rectrices when found on 15 September 1975.

I examined the stomachs of all the birds from 1973 and those found from
August-September of 1974 and found virtually all to be empty. Other workers
have recorded similar results for migrating birds (Tordoff and Mengel
1956:8). I did find, however, that Yellow-billed Cuckoos (Coccyzus amen-
canus) and Black-billed Cuckoos (C. erythropthalmus) consistently had full

stomachs,

Se.v and age ratios.-Sex and age data from the WCTV tower samples
are in Table 1 and the speeies included are only those with large sample
sizes (usually ^10). The total maleifemale ratio is virtually 1:1 (50.2%
S $ ) and the 46.2% adult proportion is essentially what others have found
for adult survival rates (Ricklefs 1973:404). Individual species often do not
have these expected ratios and this may be due to real inequalities within the
populations, year-to-year variation, or to differential migration by age or
sex classes.

Murray (1966) defined 3 types of differential migration: non-overlapping
asynchronous in which all individuals of one age or sex class precede the
others (or vice versa); overlapping asynchronous in which the timing of
the age and sex classes overlap, but one peaks before the others; and

Crawford • FLORIDA TOWER KILLS

337

Table 1

Sex and Age Ratios of Birds Killed at the WCTV Tower: Autumns 1973-1975

Species

Total

examined Ad:Im cTcT:?? Ad"d":Id"cr A9$:I9?

Yellow-billed Cuckoo

Coccyzus americanus

Acadian Flycatcher

Empidonax virescens

House Wren

Troglodytes aedon

Short-billed Marsh Wren

Cistothorus platensis

Gray Catbird

Dumetella carolinensis

Wood Thrush

Hylocichla mustelina

Swainson’s Thrush

Catharus ustulatus

Gray-cheeked Thrush

Catharus minimus

Veery

125

Catharus fuscescens

14 Sept. 1975”

15 Sept. 1975

Ruby-crowned Kinglet

Regulus calendula

White-eyed Vireo

Vireo griseus

\ ellow-throated Vireo

Vireo flavifrons

Red-eyed Vireo

896

Vireo olivaceus

5 Sept. 1974

14 Sept. 1975

251

15 Sept. 1975

237

23 Sept. 1974

17 Oct. 1974

Philadelphia Vireo

Vireo philadelphicus

17 Oct. 1974

Black-and-white Warbler

Mniotilta varia

Prothonotary Warbler

Protonotaria citrea

15 Sept. 1975

Swainson’s Warbler 8

I Am no t hi y pis s wai nsoni i

5:7

— : —

12:4^

4:10

2:1

9:1

14:24

16:15

7:9

5:10

1:19

4:9

0:4

1:8

39:50

48:34

24:24

12:22

9:17

11:11

2:9

7:4

20:56

42:15

11:31

5:10

7:12

6:8

2:4

3:5

73:46

61:58

35:22

37:19

38:30

29:37

16:13

22:15

12:10

15:10

5:7

6:2

—

9:24

— : —

—

17:7

11:12

6:3

8:3

0:10

3:5

0:3

0:5

452:444

370:357

183:187

244:113

15:39

24:22

7:17

7:15

168:83

107:112

75:32

90:22

123:114

94:99

52:42

61:38

40:20

26:24

14:12

21:3

8:90

29:20

0:29

8:12

2:4

3:3

2:1

0:3

2:4

3:3

2:1

0:3

29 :59

44:32

14:29

13:17

9:16

11:11

5:6

4:7

5:9

8:6

4:4

1:5

7:1

4:4

3:1

4:0

338

THE WILSON BULLETIN • Vol. 90, No. 3, September 1978

Table 1 (continued)

Total

Species

examined

Ad:Im

c^cf:?9

Add-.ldd

Worm-eating Warbler

7:19

5:15

3:2

3:12

Helmitheros vermivorus

Tennessee Warbler

12:44

21:20

5:16

6:14

Vermivora peregrina

11 Oct. 1974

8:21

11:11

3:8

4:7

Northern Parula

54:29

36:20

26:10

12:8

Parula arnericana

14 Sept. 1975

13:4

8:5

7:1

4:1

17 Oct. 1974

19:12

13:11

10:3

6:5

Yellow Warbler

4:14

8:5

2:6

2:3

Dendroica petechia

Magnolia Warbler

140

73:67

47:52

25:22

32:20

Dendroica magnolia

17 Oct. 1974

106

52:54

37:43

20:17

25:18

Black-throated Blue Warbler

2:11

9:3

2:7

0:3

Dendroica caerulescens

Yellow-rumped Warbler

30:24

18:22

11:6

12:9

Dendroica coronata

Cerulean Warbler

7:10

6:5

4:2

3:2

Dendroica cerulea

Blackburnian Warbler

23:75

29:32

10:19

9:23

Dendroica fusca

14 Sept. 1975

7:15

14:6

6:8

1:5

17 Oct. 1974

2:14

1:11

0:1

1:10

Chestnut-sided Warbler

45:38

33:26

18:15

16:10

Dendroica pensylvanica

17 Oct. 1974

29:26

21:23

12:8

14:9

Bay-breasted Warbler

105

69:35

26:50

19:6

34:16

Dendroica castanea

17 Oct. 1974

44:25

17:35

13:3

22:13

Prairie Warbler

10:5

5:5

3:2

4:1

Dendroica discolor

Palm Warbler

53:19

26:19

20:6

15:4

Dendroica palmarum

17 Oct. 1974

32:13

19:17

15:4

13:4

Ovenbird

219

62:157

74:84

30:44

27:57

Seiurus aurocapillus

14 Sept. 1975

13:18

12:17

6:6

7:10

15 Sept. 1975

12:30

16:15

8:8

4:11

17 Oct. 1974

105

19:86

34:42

8:26

8:34

Northern Waterthrush

21:34

22:23

10:12

10:13

Seiiirus noveboracensis

Crawford • FLORIDA TOWER KILLS

339

Table 1 (continued)

Total

Species examined Ad:Im cfd':9$ Ad'cTJcfcT A? 9:1??

Common Yellowthroat

159

47:111

55:66

21:34

23:43

Geothlypis trichas

14 Sept. 1975

3:26

6:18

1:5

2:16

17 Oct. 1974

12:20

12:13

6:6

6:7

Hooded Warbler

48:35

40:30

23:17

19:11

Wilsonia citrina
American Redstart

39:51

33:35

20:13

17:18

Setophaga ruticilla

17 Oct. 1974

19:20

14:19

8:6

11:8

Bobolink

26:5

12:12

10:1

9:3

Dolichonyx oryzivorus
Northern Oriole

4:7

2:7

1:1

3:6

Icterus galbula
Scarlet Tanager

8:5

7:6

4:3

4:2

Piranga olivacea
Summer Tanager

3:4

2:1

1:3

Piranga rubra
Indigo Bunting

29:18

23:18

17:6

11:6

Passerina cyanea
Savannah Sparrow

6:15

9:9

2:7

3:6

Passerculus sandwichensis
Grasshopper Sparrow

3:4

3:2

1:2

2:0

Ammodramus savannarum

Other species

177

79:89

74:80

32:35

41:35

Total

3223

1455:1694

1278:1267

617:629

672:546

Italicized ratios are significantly at variance from equality (P < .05 by x^)-
- Dated entries are from single night kills.

synchronous in which the timing and numbers of the sex and age classes
are the same. Differential migration behavior may he temporal, geographical,
or both ( King et al. 1965) .

Seven species exhibited evidence of temporal differential migration at the
WCTV tower (Table 2) and all cases involved age classes only. The sig-
nificance ( at the P < .05 level ) was determined by the variance test for
homogeneity of the binomial distribution ( Snedecor and Cochran 1967:240).
For the Red-eyed Vireo, Tennessee Warbler, and Ovenhird there was an early
peak by the adults, and for the Gray Catbird, Wood Thrush, Black-and-white
Warbler, and Common Yellowthroat the immatures peaked early. All repre-
sent an overlapping asynchronous migration pattern.

Taylor and Anderson (1973) studied autumn bird casualities at the

340

THE WILSON BULLETIN • VoL 90, No. 3, September 1978

Seasonal Variation in Age

Table 2
Classes at the

WCTV

Tower, Autumns

1973-1975^

August

September

October

November

Species

16-31

1-15

16-30

1-15

16-31

1-15

* Gray Catbird

0:0

0:5

9:14

18:15

12:15

0:0

* Wood Thrush

0:0

0:4

2:3

7:10

0:0

** Black-and-white Warbler

1:2

9:39

2:3

4:1

13:14

0:0

* Tennessee Warbler

0:0

2:2

3:5

7:34

0:3

Ovenbird

0:0

35:53

2:9

6:4

19:89

1:2

** Common Yellowthroat

0:0

4:42

11:31

17:17

14:20

1:2

Red-eyed Vireo
** Ad^^:Im,^5

1:3

146:98

33:30

3:26

0:30

0:0

** Ad2$:Im$$

0:2

170:72

60:14

6:11

8:14

0:0

Ad?:Im?

0:0

15:68

9:25

0:0

1:51

0:0

** Ad:Im

1:5

331:238

102:69

9:37

9:95

0:0

1 Unless otherwise noted, numbers are adults :immatures.
* Not si<^nificant (P > .05).

** Significant (P < .05).

WDBO tower in Orange County ( central peninsular ) , Florida for the years
1969-1971, and their data seem an appropriate comparison with those from
the WCTV tower for determining evidence of geographic variation in dif-
ferential migration. Table 3 compares those species from WCTV and WDBO
which had age or sex ratios significantly different from equality by simple
chi-square tests. Intertower comparisons were based on the arcsin test for
the equality of 2 percentages ( Sokal and Rohlf 1969:608). The species in i'
Table 3 are of 2 classes: (1) those that have age or sex classes significantly
at variance from equality the same way at both towers, and (2) those that
show a difference in ratios between the towers. The Acadian Flycatcher is an
exception that fits neither class. Of those species in the first class, the
Swainson’s Thrush, Tennessee Warbler, and Blackburnian Warbler show a
preponderance of immature birds at both towers, while the Veery, VTiite-
eyed Vireo, Swainson’s Warbler, Northern Parula, Palm Warbler, and
Bobolink show a preponderance of adults at both sites. The Bay-breasted
Warbler and the Ruby-crowned Kinglet had high percentages of females, while
the Northern Parula had a preponderance of males. The species making up
the second class had different age ratios at the 2 towers and all but the
Red-eyed Vireo and the Yellow-rumped Warbler had a higher percentage of
immature birds at the WCTV tower. The exceptional species ( Acadian Fly-
catcher) had a preponderance of adults at the WCTV tower but none of this
species was recorded at the WDBO site. i

Crawford • FLORIDA TOWER KILLS

341

Table 3

Species Showing Significant Differences from Equality in Age or Sex Classes at
THE WCTV Tower (This Study) and at the WDBO Tower (Taylor and Anderson

1973)

AdiJts:Imm.’s cf cf : 9 ?

Species

WCTV

WDBO

WCTV

WBDO

Acadian Flycatcher

i2:T

0:0